Class-Incremental Learning (CIL) enables models to continuously integrate new knowledge while mitigating catastrophic forgetting. Driven by the remarkable generalization of CLIP, leveraging pre-trained vision-language models has become a dominant paradigm in CIL. However, current work primarily focuses on aligning global image embeddings (i.e., [CLS] token) with their corresponding text prompts (i.e., [EOS] token). Despite their good performance, we find that they discard the rich patch-level semantic information inherent in CLIP's encoders. For instance, when recognizing a rabbit, local patches may encode its distinctive cues, such as long ears and a fluffy tail, which can provide complementary evidence for recognition. Based on the above observation, we propose SPA (Semantic-guided Patch-level Alignment) for CLIP-based CIL, which aims to awaken long-neglected local representations within CLIP. Specifically, for each class, we first construct representative and diverse visual samples and feed them to GPT-5 as visual guidance to generate class-wise semantic descriptions. These descriptions are used to guide the selection of discriminative patch-level visual features. Building upon these selected patches, we further employ optimal transport to align selected patch tokens with semantic tokens from class-wise descriptions, yielding a structured cross-modal alignment that improves recognition. Furthermore, we introduce task-specific projectors for effective adaptation to downstream incremental tasks, and sample pseudo-features from stored class-wise Gaussian statistics to calibrate old-class representations, thereby mitigating catastrophic forgetting. Extensive experiments demonstrate that SPA achieves state-of-the-art performance.

Modeling long-range dependencies in sequential data remains a central challenge in machine learning. Transformers address this challenge through attention mechanisms, but their quadratic complexity with respect to sequence length limits scalability to long contexts. State-space models (SSMs) provide an efficient alternative with linear-time computation by evolving a latent state through recurrent updates, but their memory is typically formed via additive or linear transitions, which can limit their ability to capture complex global interactions across tokens. In this work, we introduce one of the first studies to leverage the superposition property of quantum systems to enhance state-based sequence modeling. In particular, we propose Quantum Long-Attention Memory (QLAM), a hybrid quantum-classical memory mechanism that can be viewed as a quantum extension of state-space models. Instead of maintaining a classical latent state updated through additive dynamics, QLAM represents the hidden state as a quantum state whose amplitudes encode a superposition of historical information. The state evolves through parameterized quantum circuits conditioned on the input, enabling a non-classical, globally update mechanism. In this way, QLAM preserves the recurrent and linear-time structure of SSMs while fundamentally enriching the memory representation through quantum superposition. Unlike attention mechanisms that explicitly compute pairwise interactions, QLAM implicitly captures global dependencies through the evolution of the quantum state, and retrieves task-relevant information via query-dependent measurements. We evaluate QLAM on sequential variants of standard image classification benchmarks, including sMNIST, sFashion-MNIST, and sCIFAR-10, where images are flattened into token sequences. Across all tasks, QLAM consistently improves over recurrent baselines and transformer-based models.

Long-context modeling is becoming a core capability of modern large vision-language models (LVLMs), enabling sustained context management across long-document understanding, video analysis, and multi-turn tool use in agentic workflows. Yet practical training recipes remain insufficiently explored, particularly for designing and balancing long-context data mixtures. In this work, we present a systematic study of long-context continued pre-training for LVLMs, extending a 7B model from 32K to 128K context with extensive ablations on long-document data. We first show that long-document VQA is substantially more effective than OCR transcription. Building on this observation, our ablations further yield three key findings: i) for sequence-length distribution, balanced data outperforms target-length-focused data (e.g., 128K), suggesting that long-context ability requires generalizable key-information retrieval across various lengths and positions; ii) retrieval remains the primary bottleneck, favoring retrieval-heavy mixtures with modest reasoning data for task diversity; and iii) pure long-document VQA largely preserves short-context capabilities, suggesting that instruction-formatted long data reduces the need for short-data mixing. Based on these findings, we introduce MMProLong, obtained by long-context continued pre-training from Qwen2.5-VL-7B with only a 5B-token budget. MMProLong improves long-document VQA scores by 7.1% and maintains strong performance at 256K and 512K contexts beyond its 128K training window, without additional training. It further generalizes to webpage-based multimodal needle retrieval, long-context vision-text compression, and long-video understanding without task-specific supervision. Overall, our study establishes a practical LongPT recipe and an empirical foundation for advancing long-context vision-language models.

Frontier LLMs are increasingly deployed as agents that pick the next action after a long log of prior tool calls produced by the same or a different model. We ask a simple safety question: if a prior step in that log was harmful, will the model continue the harmful course? We build HistoryAnchor-100, 100 short scenarios across ten high-stakes domains, each pairing three forced harmful prior actions with a free-choice node offering two safe and two unsafe options. Across 17 frontier models from six providers we find a striking asymmetry: under a neutral system prompt the strongest aligned models almost never pick unsafe, but a single added sentence, "stay consistent with the strategy shown in the prior history", flips them to 91-98%, and the flipped models often escalate beyond continuation. Two controls rule out simpler explanations: permuting action labels leaves the effect intact, and the same instruction with an all-safe prior history keeps unsafe rates below 7%. Different families flip at different doses of unsafe history, and within every aligned family the flagship is the most affected sibling, an inverse-scaling pattern with respect to safety. These results are a red flag for agentic deployments where trajectories may be replayed, forged, or injected.

LiDAR scene generation is increasingly important for scalable simulation and synthetic data creation, especially under diverse sensing conditions that are costly to capture at scale. Typically, diffusion-based LiDAR generators are developed under single-domain settings, requiring separate models for different datasets or sensing conditions and hindering unified, controllable synthesis under heterogeneous distribution shifts. To this end, we present OmniLiDAR, a unified text-conditioned diffusion framework that generates LiDAR scans in a shared range-image representation across eight representative domains spanning three shift types: adverse weather, sensor-configuration changes (e.g., reduced beams), and cross-platform acquisition (vehicle, drone, and quadruped). To enable training a single model over heterogeneous domains without isolating optimization by domain, we introduce a Cross-Domain Training Strategy (CDTS) that mixes domains within each mini-batch and leverages conditioning to steer generation. We further propose Cross-Domain Feature Modeling (CDFM), which captures directional dependencies along azimuth and elevation axes to reflect the anisotropic scanning structure of range images, and Domain-Adaptive Feature Scaling (DAFS) as a lightweight modulation to account for structured domain-dependent feature shifts during denoising. In the absence of a public consolidated benchmark, we construct an 8-domain dataset by combining real-world scans with physically based weather simulation and systematic beam reduction while following official splits. Extensive experiments demonstrate strong generation fidelity and consistent gains in downstream use cases, including generative data augmentation for LiDAR semantic segmentation and 3D object detection, as well as robustness evaluation under corruptions, with consistent benefits in limited-label regimes.

Automated CT triage requires models that are simultaneously accurate across diverse pathologies and reliable under institutional shift. While Vision Transformers provide strong visual representations, many clinically significant findings are defined by quantitative imaging biomarkers rather than appearance alone. We introduce JANUS, a physiology-guided dual-stream architecture that conditions visual embeddings on macro-radiomic priors via Anatomically Guided Gating. On the MERLIN test set (N=5082), JANUS attains macro-AUROC 0.88 and AUPRC 0.74, outperforming all reproduced baselines. It generalizes to an external dataset N=2000; AUROC 0.87), with the largest gains on findings defined by size and attenuation as well as improved calibration on both datasets. We further quantify prediction suppression using the Physiological Veto Rate (PVR), showing that under domain shift JANUS reduces high-confidence false positives substantially more often than true positives. Together, these results are consistent with physically grounded conditioning that improves both discrimination and reliability in CT triage. Code is made publicly available at github repository https://github.com/lavsendahal/janus and model weights are at https://huggingface.co/lavsendahal/janus.

Video temporal grounding (VTG) takes an untrimmed video and a natural-language query as input and localizes the temporal moment that best matches the query. Existing methods rely on large, task-specific datasets requiring costly manual annotation. We introduce EvoGround, a framework of two coupled self-evolving agents, a proposer and a solver, that learn temporal grounding from raw videos without any human-labeled data. The proposer generates query--moment pairs from raw videos, while the solver learns to ground them and feeds back signals that improve the proposer in return. Through this self-reinforcing reinforcement-learning loop, the two agents are initialized from the same backbone and mutually improve across iterations. Trained on 2.5K unlabeled videos, EvoGround matches or surpasses fully supervised models across multiple VTG benchmarks, while emerging as a state-of-the-art fine-grained video captioner without manual labels.

Cross-modal 3D medical image analysis requires voxelwise representations that remain anatomically consistent across imaging contrasts, scanners, and acquisition protocols. Recent work has shown that frozen 2D Vision Transformer (ViT) foundation models can support such representations, but typical pipelines extract features along a single anatomical axis and adapt those features inside a registration solver for one image pair at a time, leaving complementary viewing directions unused and producing representations that do not transfer to new volumes. We introduce VoxCor, a training-free fit--transform method for reusable volumetric feature representations from frozen 2D ViT foundation models. During an offline fitting phase, VoxCor combines triplanar ViT inference with a compact closed-form weighted partial least squares (WPLS) projection that uses fitting-time voxel correspondences to select modality-stable anatomical directions in the triplanar feature space. At transform time, new volumes are mapped by triplanar ViT inference and linear projection alone, without fine-tuning or registration. Voxel correspondences can then be queried directly by nearest-neighbor search. We evaluate VoxCor on intra-subject Abdomen MR--CT and inter-subject HCP T2w--T1w tasks using deformable registration, voxelwise k-nearest-neighbor segmentation, and segmentation-center landmark localization. VoxCor improves the hardest cross-subject, cross-modality transfer settings, reduces encoder sensitivity for dense correspondence transfer, and yields registration performance competitive with handcrafted descriptors and learned 3D features. This positions VoxCor as a reusable feature layer for downstream multimodal analysis beyond pairwise registration. Code, configuration files, and implementation details are publicly available on GitHub at \href{https://github.com/guneytombak/VoxCor}{guneytombak/VoxCor}.

We present BlitzGS, a distributed 3DGS framework that reduces active Gaussian workload for fast city-scale reconstruction. BlitzGS manages this workload at three coupled levels. At the system level, the framework shards Gaussians across GPUs by index parity rather than spatial blocks. This approach mitigates the cross-block visibility redundancy inherent in spatial partitioning. Furthermore, it distributes each rendering step through a single cross-GPU exchange that routes projected Gaussians to their tile owners. At the model level, scheduled importance-scoring passes shrink the global Gaussian population. During these passes, the framework generates a per-Gaussian visibility weight to bias density-control updates toward contributing primitives and a per-view importance mask for the view-level renderer. At the view level, BlitzGS trims each camera's active set with a distance-based LOD gate to exclude excessively fine primitives for the current frustum and the importance-based culling mask to skip Gaussians with negligible cross-view contribution. On large-scale benchmarks, BlitzGS matches the rendering quality of recent large-scale baselines while delivering an order-of-magnitude speedup, training city-scale scenes in tens of minutes. Our code is available at https: //github.com/AkierRaee/BlitzGS.

Diffusion-based vision-language-action models (dVLAs) are promising for embodied intelligence but are fundamentally limited in real-time deployment by the high latency of full inference. We propose Realtime-VLA FLASH, a speculative inference framework that eliminates most full inference calls during replanning by introducing a lightweight draft model with parallel verification via the main model's Action Expert and a phase-aware fallback mechanism that reverts to the full inference pipeline when needed. This design enables low-latency, high-frequency replanning without sacrificing reliability. Experiments show that on LIBERO, FLASH largely preserves task performance by replacing many 58.0 ms full-inference rounds with speculative rounds as fast as 7.8 ms, lowering task-level average inference latency to 19.1 ms (3.04x speedup). We additionally demonstrate effectiveness on real-world conveyor-belt sorting, highlighting its practical impact for latency-critical embodied tasks.

The scalability of robotic manipulation is fundamentally bottlenecked by the scarcity of task-aligned physical interaction data. While vision-language models (VLMs) and video generation models (VGMs) hold promise for autonomous data synthesis, they suffer from semantic-spatial misalignment and physical hallucinations, respectively. To bridge this gap, we introduce RoboEvolve, a novel framework that couples a VLM planner and a VGM simulator into a mutually reinforcing co-evolutionary loop. Operating purely on unlabeled seed images, RoboEvolve leverages a cognitive-inspired dual-phase mechanism: (i) daytime exploration fosters physically grounded behavioral discovery through a semantic-controlled multi-granular reward, and (ii) nighttime consolidation mines "near-miss" failures to stabilize policy optimization. Guided by an autonomous progressive curriculum, the system naturally scales from simple atomic actions to complex tasks. Extensive experiments demonstrate that RoboEvolve (I) achieves superior effectiveness, elevating base planners by 30 absolute points and amplifying simulator success by 48% on average; (II) exhibits extreme data efficiency, surpassing fully supervised baselines with merely 500 unlabeled seeds--a 50x reduction; and (III) demonstrates robust continual learning without catastrophic forgetting.

In recent years, autonomous driving has significantly in creased the demand for high-quality data to train 2D and 3D perception models for safety-critical scenarios. Real world datasets struggle to meet this demand as require ments continuously evolve and large-scale annotated data collection remains costly and time-consuming making syn thetic data a scalable, practical and controllable alterna tive. Pedestrian detection is among the most safety-critical tasks in autonomous driving. In this paper, we propose a simple yet effective method for scaling variability in 3D pedestrian assets for synthetic scene generation. Starting from a single 3D base asset, we generate multiple distinct pedestrian instances by synthesizing diverse facial textures and identity-level appearance variations using StyleGAN2 and automatically mapping them onto 3D meshes. This ap proach enables scalable appearance-level asset diversifica tion without requiring the design of new geometries for each instance. Using the assets, we construct synthetic datasets and study the impact of mixing real and synthetic data for RGB-based object detection. Through complementary ex periments, we analyze geometry-driven distribution shifts in point cloud perception for 3D object detection. Our findings demonstrate that controlled synthetic diversifica tion improves robustness in 2D detection while revealing the sensitivity of 3D perception models to geometric domain gaps. Overall, this work highlights how generative AI en ables scalable, simulation-ready pedestrian diversification through controlled facial texture synthesis, along with the benefits and limitations of cross-domain training strategies in autonomous driving pipelines.

We propose min Generalized Sliced Gromov--Wasserstein (min-GSGW), a sliced formulation for the Gromov--Wasserstein (GW) problem using expressive generalized slicers. The key idea is to learn coupled nonlinear slicers that assign compatible push-forward values to both input measures, so that monotone coupling in the projected domain lifts to a transport plan evaluated against the GW objective in the original spaces. The resulting plan induces a GW objective value, and min-GSGW minimizes this cost directly in the original spaces. We further show that min-GSGW is rigid-motion invariant, a crucial property for geometric matching and shape analysis tasks. Our contributions are threefold: 1) we introduce generalized slicers into the sliced GW framework, 2) we construct a slicing-based efficient GW transport plan; and 3) we develop an amortized variant that replaces per-instance optimization with a learned slicer for unseen input pairs. We perform experiments on animal mesh matching, horse mesh interpolation, and ShapeNet part transfer. Results show that min-GSGW produces meaningful geometric correspondences and GW objective values at substantially lower computational cost than existing GW solvers.

In this paper, we explore a weakly supervised method for anomaly detection. Since annotating videos is time-consuming, we only look at weak video-level labels during training. This means that given a video, we know that it is either normal or contains an anomaly, but no further annotations are used to train the network. Features are extracted from video clips that are either normal or anomalous. These features are used to determine anomaly scores for spatiotemporal regions of the clips based on a classifier and the implementation of a multiple instance ranking loss (MIL). We represent both anomalous and normal video clips as positive and negative bags, respectively, to apply MIL. Furthermore, since anomalies are usually localized to a part of a frame rather than the whole frame, we chose to explore temporal as well as spatial anomaly detection. We show our results on the UCF Crime2Local Dataset, which contains spatiotemporal annotations for a portion of the UCF Crime Dataset.

Image restoration is an inherently ill posed inverse problem. Equivariant networks that embed geometric symmetry priors can mitigate this ill posedness and improve performance. However, current understanding of the relationship between network equivariance and data symmetry remains largely heuristic. Particularly for real world data with imperfect symmetry, existing research lacks a systematic theoretical framework to quantify symmetry, select transformation groups, or evaluate model data alignment. To bridge this gap, we conduct an analysis from an optimization perspective and formalize the intrinsic relationship among data symmetry priors, model equivariance, and generalization capability. Specifically, we propose for the first time a quantifiable definition of non strict symmetry at the dataset level (rather than sample level) and use it as a constraint to formulate the restoration inverse problem. We then show that the equivariance for restoration models can be naturally derived from this inverse problems incorporated the proposed symmetry constraints, and that the equivariance error of the optimal restoration operator is strictly bounded by the data symmetry error and the discretization mesh size. Furthermore, by analyzing the network's empirical risk, we demonstrate that aligning equivariance with data symmetry optimizes the bias variance trade off, minimizing the total expected risk. Guided by these insights, we propose a Sample Adaptive Equivariant Network that uses a hypernetwork and transformation learnable equivariant convolutions to dynamically align with each sample's inherent symmetry. Extensive experiments on super resolution, denoising, and deraining validate our theoretical findings and show significant superiority over standard baselines and traditional equivariant models. Our code and supplementary material are available at https://github.com/tanfy929/SA-Conv.

Scene graphs are becoming a standard representation for robot navigation, providing hierarchical geometric and semantic scene understanding. However, most scene graph mapping methods rely on depth cameras or LiDAR sensors. In this work, we present LEXI-SG, the first dense monocular visual mapping system for open-vocabulary 3D scene graphs using only RGB camera input. Our approach exploits the semantic priors of open-vocabulary foundation models to partition the scene into rooms, deferring feed-forward reconstruction to when each room is fully observed -- enabling scalable dense mapping without sliding-window scale inconsistencies. We propose a room-based factor graph formulation to globally align room reconstructions while preserving local map consistency and naturally imposing the semantic scene graph hierarchy. Within each room, we further support open-vocabulary object segmentation and tracking. We validate LEXI-SG on indoor scenes from the Habitat-Matterport 3D and self-collected egocentric office sequences. We evaluate its performance against existing feed-forward SLAM methods, as well as established scene graphs baselines. We demonstrate improved trajectory estimation and dense reconstruction, as well as, competitive performance in open-vocabulary segmentation. LEXI-SG shows that accurate, scalable, open-vocabulary 3D scene graphs can be achieved from monocular RGB alone. Our project page and office sequences are available here: https://ori-drs.github.io/lexisg-web/.

Transthoracic echocardiography (TTE) is the first-line imaging modality for diagnosing bicuspid aortic valve (BAV), yet diagnostic performance varies with operator expertise and image quality. We developed an explainable AI model that distinguishes BAV from tricuspid aortic valves (TAV) using routinely acquired parasternal long-axis (PLAX) cine loops. A multi-backbone video ensemble was trained and evaluated using a leakage-aware, stratified outer cross-validation protocol on $N{=}90$ patient studies (48 BAV, 42 TAV). Across fixed outer splits and 10 random seeds, the calibrated stacked ensemble achieved an outer-CV F1-score of $0.907$ and recall of $0.877$. Frame-level Grad-CAM localized salient evidence to the aortic root and leaflet plane, while globally aggregated SHAP values quantified each video backbone's contribution to the stacked prediction, enabling transparent, case-level auditability. These findings indicate that PLAX-based video ensembles can support reliable BAV/TAV classification from routine echocardiographic cine loops and may facilitate earlier detection in non-specialist or resource-limited clinical settings.

Trajectory-controlled human motion generation aims to synthesize realistic human motions conditioned on both textual descriptions and spatial trajectories. However, existing methods suffer from two critical limitations: first, the conflict between text and trajectory conditions disrupts the denoising process, resulting in compromised motion quality or inaccurate trajectory following; second, the use of redundant motion representations introduces inconsistencies between motion components, leading to instability during trajectory control. To address these challenges, we propose CMC, a decoupled framework that effectively coordinates text and trajectory conditions through a divide-and-conquer strategy. CMC follows a divide-and-conquer paradigm, comprising two cascaded stages: Trajectory Control and Motion Completion. In the first stage, a diffusion model generates a simplified representation of the controlled joints under trajectory guidance, based on the given trajectories, ensuring accurate and stable trajectory following. In the second stage, a text-conditioned diffusion inpainting model generates full-body motions using the simplified representation from the first stage as partial observations. To mitigate overfitting caused by limited inpainting training data, we further introduce the Selective Inpainting Mechanism (SIM), which alternates between text-to-motion generation and motion inpainting tasks during training. Experiments on HumanML3D and KIT datasets demonstrate that CMC achieves state-of-the-art performance in control accuracy and motion quality, demonstrating its effectiveness in coordinating multimodal conditions and representations.

Few-step video generation has been significantly advanced by consistency distillation. However, the performance of consistency-distilled models often degrades as more sampling steps are allocated at test time, limiting their effectiveness for any-step video diffusion. This limitation arises because consistency distillation replaces the original probability-flow ODE trajectory with a consistency-sampling trajectory, weakening the desirable test-time scaling behavior of ODE sampling. To address this limitation, we introduce AnyFlow, the first any-step video diffusion distillation framework based on flow maps. Instead of distilling a model for only a few fixed sampling steps, AnyFlow optimizes the full ODE sampling trajectory. To this end, we shift the distillation target from endpoint consistency mapping $(z_{t}\rightarrow z_{0})$ to flow-map transition learning $(z_{t}\rightarrow z_{r})$ over arbitrary time intervals. We further propose Flow Map Backward Simulation, which decomposes a full Euler rollout into shortcut flow-map transitions, enabling efficient on-policy distillation that reduces test-time errors (i.e., discretization error in few-step sampling and exposure bias in causal generation). Extensive experiments across both bidirectional and causal architectures, at scales ranging from 1.3B to 14B parameters, demonstrate that AnyFlow achieves performance matches or surpasses consistency-based counterparts in the few-step regime, while scaling with sampling step budgets.

Volumetric Modulated Arc Therapy (VMAT) is a cornerstone of modern radiation therapy, enabling highly conformal tumor irradiation and healthy-tissue sparing. Yet, its planning solves inverse and nested optimization for multi-leaf collimators, monitor units and dose parameters, while enforcing their consistency to ensure mechanical deliverability. Nevertheless, this process often requires repeated re-optimization when treatment configurations change, resulting in substantial planning time per patient. To address these problems, we present a diffusion-driven Learning-to-Optimize (L2O) method for end-to-end VMAT planning. A distribution-matching distilled diffusion model learns a clinically feasible manifold of fluence maps, enabling their one-shot generation. On top of this, an LSTM-based L2O module learns gradient update dynamics to swiftly refine fluence maps toward prescribed dose objectives during inference. Experimental results on clinical and public prostate cancer cohorts demonstrate improved planning efficiency, flexibility, and machine deliverability over currently available end-to-end VMAT planners.

Medical segmentation foundation models such as SAM and MedSAM provide strong prompt-driven segmentation, but their image encoders are still too large for many clinical settings. Compression is also risky in medicine because a model can keep high Dice while losing boundary fidelity. We propose MedCore, a structured pruning framework for MedSAM. The main idea is to preserve two kinds of structures: structures that became important during SAM-to-MedSAM adaptation, and structures that have high boundary leverage. We identify the first type by a dual-intervention score that compares zeroing a group with resetting it to its original SAM weight. We identify the second type by boundary-aware Fisher estimation. We also introduce a boundary leverage principle, which shows that compression-induced boundary displacement is controlled by logit perturbation on the boundary divided by the logit spatial gradient. This principle explains why boundary metrics can degrade even when Dice remains high. On polyp segmentation benchmarks, MedCore reduces parameters by 60.0% and FLOPs by 58.4% while achieving Dice 0.9549, Boundary F1 0.6388, and HD95 5.14 after recovery fine-tuning. It also reaches 86.6% parameter reduction and 90.4G FLOPs with strong boundary quality. Our analysis further shows that MedSAM lies in a head-fragile boundary regime: head-pruning steps have 2.887 times larger 95th-percentile boundary leverage than MLP-pruning steps, and this logit-level effect is consistent with BF1 and HD95 degradation. Our code is available at https://github.com/cenweizhang/MedCore.

Medical image-to-image (I2I) translation enables virtual scanning, i.e. the synthesis of a target imaging modality from a source one without additional acquisitions. Despite growing interest, most proposed methods operate on 2D slices, are evaluated on isolated tasks with different experimental set-ups and lack clinical validation. The primary contribution of this work is a reproducible, standardized comparative evaluation of 3D I2I translation methods in oncological imaging, designed to standardize preprocessing, splitting, inference, and multi-level evaluation across heterogeneous clinical tasks. Within this framework, we compare seven generative models, three Generative Adversarial Networks (GANs: Pix2Pix, CycleGAN, SRGAN) and four latent generative models (Latent Diffusion Model, Latent Diffusion Model+ControlNet, Brownian Bridge, Flow Matching), across eleven datasets spanning three anatomical regions (head/neck, lung, pelvis) and four translation directions (cone-beam CT to CT, MRI to CT, CT to PET, MRI T2-weighted to T2-FLAIR), for a total of 77 experiments under uniform training, inference, and evaluation conditions. The results show that GANs outperform latent generative models across all tasks, with SRGAN achieving statistically significant superiority. Our lesion-level analysis reveals that all models struggle with small lesions and that, in CT to PET synthesis, models reproduce lesion shape more reliably than absolute uptake-related intensity. We also performed a Visual Turing test administered to 17 physicians, including 15 radiologists, which shows near-chance classification accuracy (56.7%), confirming that synthetic volumes are largely indistinguishable from real acquisitions, while exposing a dissociation between quantitative metrics and clinical preference.

Deep neural networks trained with different architectures, objectives, and datasets have been reported to converge on similar visual representations. However, what remains unknown is which visual properties models actually converge on and which factors may underlie this convergence. To address this, we decompose the object similarity structure of 162 diverse vision models into a small set of non-negative dimensions. To determine universal versus model-specific dimensions, we then estimate how often each dimension reappears across models. In contrast to model-specific dimensions, universal dimensions are more interpretable and more strongly driven by conceptual image properties, indicating the relevance of interpretability and semantic content as implicit factors driving universality across models. Differences in architecture, objective function, training data, model size, and model performance do not explain the emergence of universal dimensions. However, models with more universal dimensions also better predict macaque IT activity and human similarity judgments, suggesting that universality reflects representations relevant to biological vision. These findings have important implications for understanding the emergent representations underlying deep neural network models and their alignment with biological vision.

Accurate and efficient battery detection is increasingly important for applications in electronic waste recycling, industrial quality control, and automated sorting systems. In this paper, we present both a comprehensive benchmark and a novel method for multi-class battery detection. We systematically compare three CNN-based detectors (YOLOv8n, YOLOv8s, YOLO11n) and two transformer-based detectors (RT-DETR-L, RT-DETR-X) on a publicly available dataset of approximately 8,591 annotated images under identical experimental conditions, and further propose PaQ-RT-DETR, which introduces pattern-based dynamic query generation into RT-DETR to alleviate query activation imbalance with negligible computational overhead. Among baselines, YOLO11n achieves the best CNN-based accuracy (mAP@50: 0.779) at only 2.6M parameters, while YOLOv8n delivers the fastest inference at ~1,667 FPS. PaQ-RT-DETR-X achieves the highest overall mAP@50 of 0.782, surpassing RT-DETR-X by +2.8% with consistent per-class gains across all six battery categories including the data-scarce Bike Battery class. Our findings provide practical guidance for selecting object detection models in battery-related industrial applications.

Scene graph generation provides a compact structured representation for visual perception, but accurate and fast graph prediction from images and videos remains challenging. Recent VLM-based methods can generate scene graphs end-to-end as structured text, yet often produce long outputs with irrelevant objects and relations. We present SceneGraphVLM, a compact method for image and video scene graph generation with small visual language models. SceneGraphVLM serializes graphs in a token-efficient TOON format and trains the model in two stages: supervised fine-tuning followed by reinforcement learning with hallucination-aware rewards that balance relation coverage and precision while penalizing unsupported objects and relations. For videos, the model can optionally condition each frame on the previously generated graph, providing lightweight short-term context without tracking or post-processing. We evaluate SceneGraphVLM on PSG, PVSG, and Action Genome. With compact VLMs and vLLM-accelerated decoding, SceneGraphVLM achieves a strong quality-speed trade-off, improves precision-oriented SGG metrics while preserving reasonable recall, and generates complete scene graphs with approximately one-second latency. Code and implementation details are available at: https://github.com/markus0440/SceneGraphVLM.git.

Thermal-infrared (TIR) hyperspectral imagery (HSI) provides critical scene information for various applications. However, its practical utility is severely limited by unique sensor degradations beyond the capabilities of existing restoration methods, which are ignorant of underlying thermal physics. Here, we propose HAIR (HADAR-based Image Restoration) as a physics-driven framework for ground-based TIR-HSI restoration. HAIR utilizes the HADAR rendering equation (HRE) and combines it with the atmospheric downwelling radiative transfer equation (RTE) to model TIR-HSI using temperature, emissivity, and texture (TeX) physical triplets. This physical model leads to a TeX decompose-synthesize strategy that guarantees physical consistency and spatio-spectral noise resilience, in stark contrast to existing approaches. Moreover, our framework uses a forward-modeled atmospheric downwelling reference, along with spectral smoothness of emissivity and blackbody radiation, to enable spectral calibration and generation that would otherwise be elusive. Our extensive experiments on the outdoor DARPA Invisible Headlights dataset and in-lab FTIR measurements show that HAIR consistently outperforms state-of-the-art methods across denoising, inpainting, spectral calibration, and spectral super-resolution, establishing a benchmark in objective accuracy and visual quality.

In this paper, we propose GTA-VLA(Guide, Think, Act), an interactive Vision-Language-Action (VLA) framework that enables spatially steerable embodied reasoning by allowing users to guide robot policies with explicit visual cues. Existing VLA models learn a direct "Sense-to-Act" mapping from multimodal observations to robot actions. While effective within the training distribution, such tightly coupled policies are brittle under out-of-domain (OOD) shifts and difficult to correct when failures occur. Although recent embodied Chain-of-Thought (CoT) approaches expose intermediate reasoning, they still lack a mechanism for incorporating human spatial guidance, limiting their ability to resolve visual ambiguities or recover from mistakes. To address this gap, our framework allows users to optionally guide the policy with spatial priors, such as affordance points, boxes, and traces, which the subsequent reasoning process can directly condition on. Based on these inputs, the model generates a unified spatial-visual Chain-of-Thought that integrates external guidance with internal task planning, aligning human visual intent with autonomous decision-making. For practical deployment, we further couple the reasoning module with a lightweight reactive action head for efficient action execution. Extensive experiments demonstrate the effectiveness of our approach. On the in-domain SimplerEnv WidowX benchmark, our framework achieves a state-of-the-art 81.2% success rate. Under OOD visual shifts and spatial ambiguities, a single visual interaction substantially improves task success over existing methods, highlighting the value of interactive reasoning for failure recovery in embodied control. Details of the project can be found here: https://signalispupupu.github.io/GTA-VLA_ProjPage/

Infrared-visible object detection improves detection performance by combining complementary features from multispectral images. Existing backbone-specific and backbone-shared approaches still suffer from the problems of severe bias of modality-shared features and the insufficiency of modality-specific features. To address these issues, we propose a novel detection framework WD-FQDet that explicitly decouples modality-shared and modality-specific information from infrared and visible modalities in the new view of low- and high-frequency domains, allowing fusion strategies tailored to their frequency characteristics. Specifically, a low-frequency homogeneity alignment module is proposed to align modality-shared features across modalities via a cross-modal attention mechanism, and a high-frequency specificity retention module is proposed to preserve modality-specific features through the multi-scale gradient consistency loss. To reinforce the feature representation in the frequency domain, we propose a hybrid feature enhancement module that incorporates spatial cues. Furthermore, considering that the contributions of homogeneous and modality-specific features to object detection vary across scenarios, we propose a frequency-aware query selection module to dynamically regulate their contributions. Experimental results on the FLIR, LLVIP, and M3FD datasets demonstrate that WD-FQDet achieves state-of-the-art performance across multiple evaluation metrics.

Extended depth of field microscopy encodes axial information into a single acquisition through engineered point spread functions, but conventional and deep optics approaches are subject to degradation in scattering tissue. We introduce DeepFilters, a scattering-aware deep optics framework that jointly optimizes a parameterized pupil filter and a digital-filter-based reconstruction network through a calibrated differentiable forward model to achieve broad generalization without retraining. Incorporating empirical scattering kernels, physics-guided regularization, and a hybrid genetic-gradient initialization strategy, DeepFilters extends the PSF from 16 micron to >400 micron in clear media and enables signal recovery beyond 120 micron deep in biological tissues, validated across fixed brain slices and sea urchin embryos.

Graph convolutional networks (GCNs) are widely used for 3D hand pose estimation, where the hand skeleton is encoded as a fixed adjacency graph. We revisit whether this is the most effective way to incorporate hand topology in 2D-to-3D lifting. In this paper, we perform controlled, parameter-matched ablations on the FPHA benchmark and show that standard multi-head self-attention consistently outperforms GCN baselines. Even when the GCN is strengthened with multi-hop adjacency and matched parameter count, self-attention reduces MPJPE from 12.36 mm to 10.09 mm. A skeleton-constrained graph attention network recovers most of this gap, indicating that input-dependent aggregation is a major source of improvement, while fully connected attention yields additional gains. We further show that hand topology is most effective when introduced as a soft structural prior through graph-distance positional encoding, rather than as a hard adjacency constraint. These results suggest that, for hand pose lifting, adaptive spatial attention is a more effective inductive bias than fixed graph convolution.

3D Language Gaussian Splatting (3DLGS) augments 3D Gaussian Splatting with language-aligned visual features for open-vocabulary 3D scene understanding. A core challenge is efficiently associating high-dimensional vision-language embeddings with millions of 3D Gaussians while preserving efficient feature rendering for text-based querying. Existing methods either store dense features directly on Gaussians, causing high storage costs and slow rendering, or learn compact representations through expensive per-scene optimization with repeated feature rasterization. No existing method simultaneously achieves fast 3D semantic reconstruction, efficient storage, and fast rendering. We propose SCOUP (Sparse COde UPlifting), which addresses all three by decoupling language representation learning from 3D Gaussian optimization. Rather than working directly in 3D, we learn sparse codebook-based representations entirely using features associated with 2D image regions, associating each region with a sparse set of codebook coefficients. We then uplift these coefficients to 3D Gaussians with our weighted sparse aggregation using Gaussian-to-pixel associations, where each Gaussian accumulates coefficients over codebook atoms across views. Top-$K$ filtering then extracts the most dominant multi-view coefficients per Gaussian, enabling efficient storage and fast rendering. Our method achieves up to $400\times$ training speedup while being $3\times$ more memory efficient during training compared to the state-of-the-art in rendering speed. Across multiple benchmarks, SCOUP matches or outperforms existing methods in open-vocabulary querying accuracy.

Reliable autonomous driving relies on large-scale, well-labeled data and robust models. However, manual data collection is resource-intensive, and traditional simulation suffers from a persistent reality gap. While recent generative frameworks and radiance-field methods improve visual fidelity, they still struggle with temporal and spatial consistency and cannot ensure physics-aware behavior, limiting their applicability to driving scenario generation. To address these challenges, we propose Real2Sim, an unified framework that combines 4D Gaussian Splatting (4DGS) with a differentiable Material Point Method (MPM) solver. Real2Sim explicitly reconstructs dynamic driving scenes as temporally continuous Gaussian primitives, supports instance-level editing, and simulates realistic object-object and object-environment interactions. This framework enables physics-aware, high-fidelity synthesis of diverse, editable scenarios, including challenging corner cases such as collisions and post-impact trajectories. Experiments on the Waymo Open Dataset validate Real2Sim's capabilities in rendering, reconstruction, editing, and physics simulation, demonstrating its potential as a scalable tool for data generation in downstream tasks such as perception, tracking, trajectory prediction, and end-to-end policy learning.

Generating controllable and physically plausible indoor scenes is a pivotal prerequisite for constructing high-fidelity simulation environments for embodied AI. However, existing deeplearning-based methods usually treat all objects as homogeneous instances within a unified generation process. While effective for sparse and simplistic layouts, they struggle to model realistic layouts with dense object arrangements and complex spatial dependencies, leadingto limited scalability and degraded physical plausibility. To deal with these challenges, we revisit indoor layout generation from the perspective of structural heterogeneity and decompose the objects into primary objects and secondary objects according to their distinct roles in shaping a scene. Based on this decomposition, we propose HetScene, a heterogeneous two-stage generation framework that decouples indoor layout synthesis into Structural Layout Generation (SLG) and Contextual Layout Generation (CLG). SLG first generates globally coherent structural layouts with only primary objects conditioned on text descriptions, top-down binary room masks, and spatial relation graphs, establishing a stable global macro-skeleton of large core furniture.

Recent advances have demonstrated that coded aperture snapshot spectral imaging (CASSI) systems show great potential for capturing 3D hyperspectral images (HSIs) from a single 2D measurement. Despite the inherent spectral continuity of scenes captured by CASSI, most existing reconstruction methods are restricted to fixed, discrete spectral outputs, thereby precluding continuous spectral reconstruction or spectral super-resolution. To address this challenge, we propose Phy-CoSF, which synergizes deep unfolding networks with implicit neural representations, establishing a new paradigm for continuous spectral reconstruction and super-resolution in CASSI. Specifically, we propose a two-phase architecture that bridges discrete-wavelength training with continuous spectral rendering, enabling the synthesis of high-fidelity HSIs at arbitrary target wavelengths. At the core of our framework lies the continuous spectral fields (CoSF) module, embedded within each unfolding stage as a dynamic prior, which comprises a triple-branch cross-domain feature mixer for comprehensive spatial-frequency-channel feature fusion, alongside a spectral synthesis head that generates spectral intensities by querying continuous wavelength coordinates. Extensive experimental results demonstrate that Phy-CoSF not only achieves continuous modeling at arbitrary spectral resolutions but also outperforms many state-of-the-art methods in both reconstruction fidelity and spectral detail preservation. Our code and more results are available at: https://github.com/PaiDii/Phy-CoSF.git.

Hyperspectral image (HSI) restoration is crucial for reliable analysis, as real HSIs suffer from degradations like noise, blur, and resolution loss. However, existing models trained on source data often fail on target domains lacking clean references, a common occurrence in practice. To address this issue, we present HIR-ALIGN, a plug-and-play target-adaptive augmentation framework that enhances hyperspectral image restoration by augmenting limited training images with synthetic data that closely matches the target distribution using no extra data. It consists of three stages: (i) proxy generation, where off-the-shelf restoration models restore degraded target observations to produce semantics-preserving proxy HSIs that approximate target-domain clean images; (ii) distribution-adaptive synthesis, where a blur-robust unCLIP diffusion model generates target-aligned RGBs from proxy RGBs, with prompt conditioning and embedding-space noise initialization. Then, a warp-based spectral transfer module synthesizes HSIs by aligning each generated RGB with the proxy RGB, estimating soft patch-wise transport weights, and applying these weights and learnable local interpolation kernels to the proxy HSI; and (iii) aligned supervised finetuning, where restoration networks pretrained on the source distribution are finetuned using both the proxy HSIs and synthesized target-aligned HSIs, and are then deployed on degraded target images. We further provide theoretical analysis showing that augmentation-based finetuning can achieve lower target-domain restoration risk by jointly improving target distribution coverage and controlling spectral bias. Extensive experiments on simulated and real datasets across denoising and super-resolution tasks demonstrate that HIR-ALIGN consistently improves source-only supervised baselines, outperforming both source-only counterparts and representative unsupervised methods.

We present Qwen-Image-VAE-2.0, a suite of high-compression Variational Autoencoders (VAEs) that achieve significant advances in both reconstruction fidelity and diffusability. To address the reconstruction bottlenecks of high compression, we adopt an improved architecture featuring Global Skip Connections (GSC) and expanded latent channels. Moreover, we scale training to billions of images and incorporate a synthetic rendering engine to improve performance in text-rich scenarios. To tackle the convergence challenges of high-dimensional latent space, we implement an enhanced semantic alignment strategy to make the latent space highly amenable to diffusion modeling. To optimize computational efficiency, we leverage an asymmetric and attention-free encoder-decoder backbone to minimize encoding overhead. We present a comprehensive evaluation of Qwen-Image-VAE-2.0 on public reconstruction benchmarks. To evaluate performance in text-rich scenarios, we propose OmniDoc-TokenBench, a new benchmark comprising a diverse collection of real-world documents coupled with specialized OCR-based evaluation metrics. Qwen-Image-VAE-2.0 achieves state-of-the-art reconstruction performance, demonstrating exceptional capabilities in both general domains and text-rich scenarios at high compression ratio. Furthermore, downstream DiT experiments reveal our models possess superior diffusability, significantly accelerating convergence compared to existing high-compression baselines. These establish Qwen-Image-VAE-2.0 as a leading model with high compression, superior reconstruction, and exceptional diffusability.

Fine-grained Vision-Language Pre-training (FVLP) demonstrates significant potential in 3D medical image understanding by aligning anatomy-level visual representations with corresponding textual descriptions. However, existing FVLP paradigms often suffer from severe representation collapse in the textual embedding space, where text embeddings of distinct anatomical structures become highly clustered and indistinguishable. This distributional degeneracy renders the model hypersensitive to prompt variations, hindering reliable clinical deployment. To address these challenges, we propose a novel Cross-Anatomy Global-Local Contrastive Learning framework (CA-GCL). CA-GCL introduces a global contrastive objective that enforces separation between anatomical categories in the latent space, effectively counteracting the aggregation tendency induced by local alignment. Furthermore, we incorporate a clinical-aware text augmentation strategy based on permutation invariance and partial completeness to enhance robustness against descriptive incompleteness. Extensive evaluations on the CT-RATE and Rad-ChestCT datasets demonstrate that CA-GCL consistently outperforms existing VLP paradigms in zero-shot abnormality detection, achieving superior performance while exhibiting strong cross-dataset generalization. Crucially, CA-GCL reduces performance variance across diverse prompt templates, transforming the collapsed textual similarity distribution into a bell-shaped distribution. These results validate CA-GCL as an effective framework for robust 3D medical image understanding.

Surgical scene understanding is a cornerstone of computer-assisted intervention. While recent advances, particularly in surgical image segmentation, have driven progress, real-world clinical applications require a more holistic understanding that jointly captures procedural context, semantic reasoning, and precise visual grounding. However, existing approaches typically address these components in isolation, leading to fragmented representations and limited semantic consistency. To address this limitation, we propose SurgMLLM, a unified surgical scene understanding framework that bridges high-level reasoning and low-level visual grounding within a single model. Given surgical videos, SurgMLLM fine-tunes a multimodal large language model (MLLM) to support structured interpretability reasoning, which is used to jointly model phases, instrument-verb-target (IVT) triplets, and triplet-entity segmentation tokens. These tokens are then temporally aggregated and serve as prompts for a segmentation network, enabling accurate pixel-wise grounding of triplet instruments and targets. The entire framework is trained end-to-end with a unified objective that couples language-based reasoning supervision with visual grounding losses, promoting coherent cross-task learning and clinically consistent scene representations. To facilitate unified evaluation, we introduce CholecT45-Scene, extending CholecT45 dataset with 64,299 frames of pixel-level mask annotations for instruments and targets, aligned with existing triplet labels. Extensive experiments show that SurgMLLM significantly advances surgical scene understanding, improving the primary triplet recognition metric AP_IVT from 40.7% to 46.0% and consistently outperforming prior methods in phase recognition and segmentation. These results highlight the effectiveness of unified reasoning-and-grounding for reliable, context-aware surgical assistance.

Can general-purpose image editors predict physical maps from a single RGB image? General-purpose image editors differ from standard task-specific dense-prediction models: they do not directly take an image and output a physical map. Instead, they must be guided by prompts, examples, or image-based textual cues. To this end, we introduce PhysEditBench, a novel protocol-conditioned benchmark to evaluate and standardize image editors in dense physical-map prediction that covers five targets: depth, normal, albedo, roughness, and metallic maps. For evaluation data, we build a target-dependent benchmark substrate. We use OpenRooms-FF for depth, surface normal, albedo, and roughness, InteriorVerse as an additional source for depth, normal, albedo, and a new procedurally generated source for metallic maps. We curate the data with quality checks, valid-region masks, scene-level sampling, and lighting-based stress subsets to ensure reliable and diverse evaluation. For each target, PhysEditBench defines a fixed protocol that specifies the allowed input, expected output format, and scoring procedure. Each score, therefore, reflects the performance of a model under a specified protocol, rather than its best possible performance under all prompts or interaction modes. Experimental results show that specialized models remain much stronger on depth, normal, and albedo, and stronger image editors can produce more reasonable map-like outputs. For roughness and metallic, image editors can match or outperform specialized baselines on some scalar metrics, but they still suffer from structural errors, sparsity effects, and sensitivity to lighting.

Content-adaptive compression has always been a key direction in neural video coding (NVC), aiming to mitigate the domain gap between training and testing data. Such gaps often arise from distributional discrepancies between training and inference data, which may cause noticeable performance degradation when the testing content differs from the training distribution. To tackle this challenge, we propose DCVC-DT, a domain transfer enhanced neural video compression framework. Specifically, we design a lightweight online domain transfer (DT) mechanism that dynamically adapts the encoded latent representation during inference, effectively bridging the domain gap without modifying the encoder or decoder parameters. In addition, we develop a frame-level dynamic RD (Rate and Distortion) adjustment scheme that actively regulates the ratio of R and D in the loss function based on quality fluctuation, thereby improving rate-distortion performance. Extensive experiments demonstrate that DCVC-DT achieves up to 6.21% bitrate savings over the baseline DCVC-DC, while significantly enhancing generalization to unseen testing data and alleviating error propagation. Our code is available at https://github.com/SunnyMass/DCVC-DT.

Recent advances in 3D Gaussian Splatting (3DGS) have enabled significant progress in photorealistic novel view synthesis. However, traditional 3DGS relies on a slow, iterative optimization process, which limits its use in scenarios demanding real-time results. To overcome this bottleneck, recent feed-forward methods aim to predict Gaussian attributes directly from images, but they often struggle with the redundancy of Gaussian primitives and rendering quality. In this work, we introduce a transformer-based architecture specifically designed for feed-forward Gaussian Splatting. Our key insight is that spatial and semantic relationships among Gaussians can be effectively captured through a sparse attention mechanism, enabled by a Z-order strategy that organizes the unstructured Gaussian set into a spatially coherent sequence. Furthermore, we incorporate this Z-order strategy to adaptively suppress redundancy while preserving critical structural details. This allows the transformer to efficiently model context, compress Gaussian primitives, and predict Gaussian attributes in a single forward pass. Comprehensive experiments demonstrate that our method achieves fast and high-quality novel view synthesis with fewer Gaussian primitives.

Diffusion-based real-world image super-resolution (Real-ISR) has achieved remarkable perceptual quality; however, directly super-resolving images to 4K remains limited by extreme memory consumption. Consequently, prior methods adopt patch-based inference, sacrificing global context and introducing semantic confusion, spatial inconsistency, and severe latency. We propose OP4KSR, a one-step patch-free 4K SR approach built upon the powerful Flux backbone. By leveraging the extreme-compression F16 VAE, OP4KSR makes 4K SR inference tractable under practical GPU budgets, preserving global spatial-semantic coherence while enabling highly efficient inference. However, adapting this one-step architecture intrinsically triggers severe periodic artifacts. We trace this to a RoPE base frequency allocation mismatch and intra-token spatial ambiguity, both exacerbated by the lack of iterative refinement. To suppress these artifacts, we couple RoPE base frequency rescaling (RFR) with an autocorrelation-based periodicity loss ($\mathcal{L}_\text{AP}$). Furthermore, we curate a dedicated training dataset alongside three benchmarks (one synthetic and two real-world) to advance 4K SR research. Extensive experiments demonstrate that OP4KSR achieves competitive perceptual quality with efficient inference, generating a $4096\times4096$ output in only 5.75 seconds on a single NVIDIA H20 GPU.

Latent Action Models (LAMs) have emerged as an effective paradigm for handling heterogeneous datasets during Vision-Language-Action (VLA) model pretraining, offering a unified action space across embodiments. However, existing LAMs often rely on discrete quantization encode and decode pipelines, which can lead to trivial frame reconstruction behavior, limited representational capacity, and a lack of physically meaningful structure. We introduce RotVLA, a VLA framework built on a continuous rotational latent action representation. Latent actions are modeled as elements of SO(n), providing continuity, compositionality, and structured geometry aligned with real-world action dynamics. A triplet frame learning framework further enforces meaningful temporal dynamics while avoiding degeneration. RotVLA consists of a VLM backbone and a flow-matching action head, pretrained on large-scale cross-embodiment robotic datasets and human videos with latent-action supervision. For downstream robot control, the flow-matching head is extended into a unified action expert that jointly denoises latent and robot actions. Here, latent actions serve as a latent planner, providing high-level guidance that conditions action generation. With only 1.7B parameters and 1700+ hours of pretraining data, RotVLA achieves 98.2% on LIBERO and 89.6% / 88.5% on RoboTwin2.0 under clean and randomized settings, respectively. It also demonstrates strong real-world performance on manipulation tasks, consistently outperforming existing VLA models.

Computing a minimum $s$-$t$ cut in a graph is a solution to a wide range of computer vision problems, and is often done using the Boykov-Kolmogorov (BK) algorithm. In this paper, we revisit the BK algorithm from both a theoretical and practical point of view. We improve the analysis of the time complexity of the BK algorithm to $O(mn|C|)$ and propose a new algorithm, the fast and compact BK (fcBK) algorithm, with a time complexity of $O(m|C|)$, where $m$, $n$, and $|C|$ are the number of edges, number of vertices, and the capacity of the cut, respectively. We additionally propose a compact graph representation that allows our implementation to find a minimum $s$-$t$ cut in a graph with upwards of $10^9$ vertices and $10^{10}$ edges on a machine with 128 GB of memory. We find our implementation of the BK algorithm to be the fastest available implementation of the BK algorithm when evaluating on a comprehensive set of benchmark datasets, highlighting the importance of memory-efficient implementations. We make our implementations publicly available for further research and implementation development within minimum $s$-$t$ cut algorithms.

Face Image Quality Assessment (FIQA) evaluates the utility of a face image for automated face recognition (FR) systems. In this work, we propose PreFIQs, an unsupervised and training-free FIQA framework grounded in the Pruning Identified Exemplar (PIE) hypothesis. We hypothesize that low-utility face images rely disproportionately on fragile network parameters, resulting in larger geometric displacement of their embeddings under model sparsification. Accordingly, PreFIQs quantifies image utility as the Euclidean distance between L2-normalized embeddings extracted from a pre-trained FR model and its pruned counterpart. We provide a first-order theoretical justification via a Jacobian-vector product analysis, demonstrating that this empirical drift serves as a computationally efficient approximation of the exact geometric sensitivity of the latent embedding manifold. Extensive experiments across eight benchmarks and four FR models demonstrate that PreFIQs achieves competitive or superior performance compared to state-of-the-art FIQA methods, including establishing new state-of-the-art results on several benchmarks, without any training or supervision. These results validate parameter sparsification as a principled and practically efficient signal for face image utility, and demonstrate that quality is, in essence, what survives pruning.

Deep neural networks are highly vulnerable to adversarial examples, i.e.,small perturbations that can significantly degrade model performance. While adversarial training has become the primary defense strategy, most studies focus on balanced datasets, overlooking the challenges posed by real-world long-tail data. Motivated by the fact that perturbations in adversarial examples inherently alter the training distribution, we theoretically investigate their impact. We first revisit adversarial training for long-tail data and identify two key limitations: (i) a skewed training objective caused by class imbalance, and (ii) unstable evolution of adversarial distributions. Furthermore, we show that perturbations can simultaneously address both adversarial vulnerability and class imbalance. Based on these insights, we propose RobustLT, a plug-and-play framework that adaptively adjusts perturbations during adversarial training. Extensive experiments demonstrate that RobustLT consistently enhances adversarial robustness and class-balance on long-tailed datasets. The code is available at \href{https://github.com/zhang-lilin/RobustLT}{https://github.com/zhang-lilin/RobustLT}.

As AI-generated synthetic images become increasingly realistic, Vision Transformers (ViTs) have emerged as a cornerstone of modern deepfake detection. However, the prevailing reliance on frozen, pre-trained backbones introduces a subtle yet critical vulnerability. In this work, we present the Surrogate Iterative Adversarial Attack (SIAA), a gray-box attack that exploits knowledge of the detector's ViT backbone alone and operates entirely within the target detector's feature space to craft highly effective adversarial examples. Through our experiments, involving multiple ViT-based detectors and diverse gray-box scenarios, including few-shot learning, complete training misalignment and attack transferability tests, we demonstrate that this vulnerability consistently yields high attack success rates, often approaching white-box performance. By doing so, we reveal that backbone knowledge alone is sufficient to undermine detector reliability, highlighting the urgent need for more resilient defenses in adversarial multimedia forensics.

In Vision-Language Models (VLMs), processing a massive number of visual tokens incurs prohibitive computational overhead. While recent training-aware pruning methods attempt to selectively discard redundant tokens, they largely rely on continuous-gradient relaxations. However, visual token pruning is inherently a discrete, non-convex combinatorial problem; consequently, these continuous approximations frequently trap the optimization in sub-optimal local minima, especially under aggressive compression budgets. To overcome this fundamental bottleneck, we propose GRIP-VLM, a Group-Relative Importance Pruning framework driven by Reinforcement Learning. Rather than relying on smooth-gradient assumptions, GRIP-VLM formulates pruning as a Markov Decision Process, employing a Group Relative Policy Optimization (GRPO) paradigm anchored by supervised warm-up to directly explore the discrete selection space. Integrated with a budget-aware scorer, our lightweight agent dynamically evaluates per-token importance and adapts to arbitrary compression ratios without retraining. Extensive experiments across diverse multimodal benchmarks demonstrate that GRIP-VLM consistently outperforms heuristic and supervised-learning baselines, achieving a superior Pareto frontier and delivering up to a 15\% inference speedup at equal accuracy.

Accurate forward modelling is essential for non-invasive cardiac electrophysiology, particularly in atrial fibrillation, where electrical activation is highly disorganised. Conventional physics-based forward models require explicit specification of intracellular conductivity tensors, which are not directly measurable in clinical practice and introduce structural modelling errors. This proof-of-concept study presents a deep learning approach that learns a direct mapping from left atrial intracellular electrical potentials to far-field ECGs without requiring explicit intracellular conductivity inputs at inference time. Despite training only on 74 subjects, the model achieved an R2 of 0.949 \pm 0.037, highlighting potential to reduce structural uncertainty and improve non-invasive AF assessment.

Diffusion-based point editing methods have gained significant traction in image editing tasks due to their ability to manipulate image semantics and fine details by applying localized perturbations on the manifold of noise latent. However, these approaches face several limitations. Traditional point-based editing relies on pairs of handle and target points to define motion trajectories, which can introduce ambiguity or unnecessary alterations. Furthermore, when the distance between the handle and target points is large, the accumulated perturbations often cause the noise latent deviation from inversion score trajectory, resulting in unnatural artifacts. To address these issues in global editing tasks, we introduce a CLIP-based model to evaluate and guide intermediate editing steps, ensuring that the generated results remain both semantically aligned. Additionally, we propose a prior-preservation loss that constrains the optimized latent code to stay within the sampling space of the diffusion prior, improving consistency with the original data distribution, to ensure the model generates images along a familiar score trajectory. For fine-grained tasks, we present a directionally-weighted point tracking mechanism that steers the editing process toward the target direction within similar feature regions. This improves both the tracking accuracy and generation quality, while also reducing the editing time.

Multi-reference image generation aims to synthesize images from textual instructions while faithfully preserving subject identities from multiple reference images. Existing VLM-enhanced diffusion models commonly rely on decoupled visual conditioning: semantic ViT features are processed by the VLM for instruction understanding, whereas appearance-rich VAE features are injected later into the diffusion backbone. Despite its intuitive design, this separation makes it difficult for the model to associate each semantically grounded subject with visual details from the correct reference image. As a result, the model may recognize which subject is being referred to, but fail to preserve its identity and fine-grained appearance, leading to attribute leakage and cross-reference confusion in complex multi-reference settings. To address this issue, we propose UniCustom, a unified visual conditioning framework that fuses ViT and VAE features before VLM encoding. This early fusion exposes the VLM to both semantic cues and appearance-rich details, enabling its hidden states to jointly encode the referred subject and corresponding visual appearance with only a lightweight linear fusion layer. To learn such unified representations, we adopt a two-stage training strategy: reconstruction-oriented pretraining that preserves reference-specific appearance details in the fused hidden states, followed by supervised finetuning on single- and multi-reference generation tasks. We further introduce a slot-wise binding regularization that encourages each image slot to preserve low-level details of its corresponding reference, thereby reducing cross-reference entanglement. Experiments on two multi-reference generation benchmarks demonstrate that UniCustom consistently improves subject consistency, instruction following, and compositional fidelity over strong baselines.

Dropout-based sparse-view 3D Gaussian Splatting (3DGS) methods alleviate overfitting by randomly suppressing Gaussian primitives during training. Existing methods mainly focus on designing increasingly sophisticated dropout strategies, while they overlook the resulting inconsistencies among different dropped Gaussian subsets. This oversight often leads to unstable reconstruction and suboptimal Gaussian representation learning.In this paper, we revisit dropout-based sparse-view 3DGS from a consistency regularization perspective and propose PairDropGS, a Paired Dropout-induced Consistency Regularization framework for sparse-view Gaussian splatting. Specifically, PairDropGS first constructs a pair of the dropped Gaussian subsets from a shared Gaussian field and designs a low-frequency consistency regularization to constrain their low-frequency rendered structures. This design encourages the shared Gaussian field to preserve stable scene layout and coarse geometry under different random dropouts, while avoiding excessive constraints on ambiguous high-frequency details. Moreover, we introduce a progressive consistency scheduling strategy to gradually strengthen the consistency regularization during training for stability and robustness of reconstruction. Extensive experiments on widely-used sparse-view benchmarks demonstrate that PairDropGS achieves superior training stability, significantly outperforms existing dropout-based 3DGS methods in reconstruction quality, while exhibiting the simplicity and plug-and-play nature for improving dropout-based optimization.

As large language models span dense, mixture-of-experts, and state-space architectures and are deployed on heterogeneous accelerators under increasingly diverse multimodal workloads, optimising inference energy has become as critical as optimizing latency and throughput. Existing approaches either treat latency as an energy proxy or rely on data-hungry black-box surrogates. Both fail under varying parallelism strategies: latency and energy optima diverge in over 20% of configurations we tested, and black-box surrogates require hundreds of profiling samples to generalize across model families and hardware. We present EnergyLens, which uses symbolic regression as a structure-discovery tool over profiling data to derive a single twelve-parameter closed-form energy model expressed in terms of system properties such as degree of parallelism, batch size, and sequence length. Unlike black-box surrogates, EnergyLens decouples tensor and pipeline parallelism contributions and separates prefill from decode energy, making its predictions physically interpretable and actionable. Fitted from as few as 50 profiling measurements, EnergyLens achieves 88.2% Top-1 configuration selection accuracy across many evaluation scenarios compared to 60.9% for the closest prior analytical baseline, matches the predictive accuracy of ensemble ML methods with 10x fewer profiling samples, and extrapolates reliably to unseen batch sizes and hardware platforms without structural modification, making it a practical, interpretable tool for energy-optimal LLM deployment.

This paper presents a novel Direct Integration Theorem (DIT), derived as a non-trivial corollary of the classical Central Slice Theorem (CST). The DIT provides a mathematically consistent transition from the continuous to the discrete domain - a fundamental challenge in computed tomography - thereby eliminating the need for frequency-domain interpolation without resorting to conventional ramp-filtering. The proposed approach circumvents two principal limitations inherent in traditional methods: (i) the zero-frequency singularity and spectral distortions introduced by the mandatory ramp-filtering step, and (ii) discretization inaccuracies associated with frequency-domain interpolation. Based on the DIT, we develop a rigorous framework for consistent discrete solutions of the inverse Radon problem. Mathematical modeling demonstrates that this approach achieves quasi-exact reconstruction, with errors constrained solely by sampling parameters and grid geometry. Furthermore, while Filtered Back Projection (FBP) inherently distorts the variance of the reconstructed image, the DIT-based algorithm preserves it. Comparative simulations confirm that the proposed method eliminates common artifacts, such as intensity cupping, and consistently outperforms FBP in terms of PSNR, SSIM, and reprojection fidelity, faithfully restoring the original image's statistical characteristics.

Underwater environments impose severe constraints on conventional imaging systems and demand solutions that balance high-quality sensing with strict resource efficiency. While emerging event cameras offer a promising alternative, their potential in aquatic scenarios remains largely unexplored. Through the lens of neuromorphic vision, this work pioneers the investigation of motion fields that serve as key media for agile underwater perception. Built upon spiking neural networks, we introduce a self-supervised framework to estimate per-pixel optical flow from asynchronous event streams, elegantly bypassing the long-standing bottleneck of underwater data scarcity. Extensive evaluations demonstrate that our method achieves competitive visual and quantitative results against leading techniques while operating with superior computational efficiency. By bridging neuromorphic sensing and aquatic intelligence, this work opens new frontiers for lightweight, real-time, and low-cost perception on resource-constrained underwater edge platforms.

We present 3D Surface Splatting (3DSS), the first differentiable surface splatting renderer for physically-based inverse rendering from multi-view images. Our central insight is that the surface separation problem at the heart of surface splatting admits a direct formulation in terms of the reconstruction kernels themselves. From this foundation we derive a coverage-based compositing model whose per-layer opacity arises directly from the accumulated Elliptical Weighted Average reconstruction weight, yielding anti-aliased silhouettes and informative visibility gradients at sparsely covered edges. Combined with forward microfacet shading under co-optimized HDR environment lighting and density-aware adaptive refinement, 3DSS jointly recovers shape, spatially-varying BRDF materials, and illumination. Because the optimized representation is a set of oriented surface samples, it bridges natively to mesh-based workflows via surface reconstruction from oriented point cloud methods. We evaluate 3DSS against mesh-based, implicit, and Gaussian-splatting baselines across geometry reconstruction, novel-view synthesis, and novel-illumination relighting.

High-resolution satellite images are often scarce and costly, especially for remote areas or infrequent events. This shortage hampers the development and testing of machine learning models for land-cover classification, change detection, and disaster monitoring. In this paper, we tackle the problem of geometry-controlled high-resolution satellite image synthesis by adding control over existing pre-trained diffusion models. We propose a simple yet efficient method for controlling the synthesis process by leveraging only skip connection features using windowed cross-attention modules. Several previously established control techniques are compared, indicating that our method achieves comparable performance while leading to a better alignment with the geometry control map. We also discuss the limitations in current evaluation approaches, amplifying the necessity of a consistent alignment assessment.

Open-world text-guided class-agnostic counting (CAC) has emerged as a flexible paradigm for counting arbitrary object classes by using natural language prompts. However, current evaluation protocols primarily focus on standard counting errors within single-category images, overlooking a fundamental requirement: the ability to correctly ground the textual prompt in the visual scene. In this paper, we show that several state-of-the-art CAC models often struggle to determine which object class should be counted based on the given prompt, revealing a misalignment between textual semantics and visual object representations. This limitation leads to spurious counting responses and reduced reliability in real-world scenarios. To systematically address these limitations, we propose a new evaluation framework focused on model robustness and trustworthiness. Our contribution is two-fold: (i) we introduce PrACo++ (Prompt-Aware Counting++), a novel test suite featuring two dedicated evaluation protocols -- the negative-label test and the distractor test -- paired with new specialized metrics; and (ii) we present the MUCCA (MUlti-Category Class-Agnostic counting) evaluation dataset, a new collection of real-world images featuring multiple annotated object categories per scene, unlike existing CAC benchmarks that typically include a single category per image. Our extensive experimental evaluation of 10 state-of-the-art methods shows that, despite strong performance under standard counting metrics, current models exhibit significant weaknesses in understanding and grounding object class descriptions. Finally, we provide a quantitative analysis of how semantic similarity between prompts influences these failures. Overall, our results underscore the need for more semantically grounded architectures and offer a reliable framework for future assessment in open-world text-guided CAC methods.

Affective Image Editing (AIE) aims to modify visual content to evoke targeted emotions. Although current approaches achieve impressive editing quality, they often overlook inference efficiency, which limits their applicability in computational social scenarios. Moreover, most methods depend on discrete emotion representations, which hinder the continuous modeling of complex human emotions and constrain expressive capabilities in interactive scenarios. To tackle these gaps, we propose MooD, the first framework that directly leverages continuous Valence-Arousal (VA) values as editing instruction for fine-grained and efficient AIE in computational social systems. Specifically, we first introduce a VA-Aware retrieval strategy to bridge vague affective values and detailed visual semantics. Building upon this, MooD integrates visual transfer and perception-enhanced semantic guidance to achieve controllable AIE. Furthermore, considering that existing VA-annotated datasets mainly focus on social scenarios and largely overlook natural scenes, we therefore construct AffectSet, a comprehensive VA-annotated dataset covering diverse scenarios, to support model optimization and evaluation. Extensive qualitative and quantitative experimental results demonstrate that our MooD achieves superior performance in both affective controllability and visual fidelity while maintaining high efficiency. A series of ablation studies further reveal the crucial factors of our design.

Scalable compression is essential for bandwidth-adaptive transmission, yet most learned codecs are optimized for a fixed rate-distortion point, making rate adaptation costly due to re-encoding or maintaining multiple bitstreams. In this work, we propose TAFA-GSGC, a scalable learned point cloud geometry codec that enables multi-quality decoding from a single bitstream and a single trained model. TAFA-GSGC combines layered residual refinement with channel-group entropy coding, and introduces a Target-Aligned Feature Aggregation module to reduce cross-layer redundancy in enhancement residuals. Our framework supports up to 9 decodable quality levels with monotonic quality improvement as more subbitstreams are received, while maintaining strong compression efficiency. Compared with the PCGCv2 baseline, TAFA-GSGC demonstrates improved RD performance, achieving average BD-rate reductions of 4.99% and 5.92% in terms of D1-PSNR and D2-PSNR, respectively.

Web-scale 3D asset collections are abundant but rarely deployment-ready, suffering from arbitrary metric scaling, incorrect pivots, brittle geometry, and incomplete textures, defects that limit their use in embodied AI, robotics, and spatial computing. We present AmaraSpatial-10K, a dataset of over 10,000 synthetic 3D assets optimised for zero-shot deployment. Each asset ships as a metric-scaled, deterministically anchored .glb with separated PBR maps, a convex collision hull, a paired reference image, and multi-sentence text metadata. Alongside the dataset we introduce a reusable evaluation suite for 3D asset banks, a continuous Scale Plausibility Score (SPS), an LLM Concept Density metric, anchor-error auditing, and a cross-modal CLIP coherence protocol, and apply it to AmaraSpatial-10K alongside matched subsets of Objaverse, HSSD, ABO, and GSO. AmaraSpatial-10K improves CLIP Recall@5 by $3.4\times$ over Objaverse ($0.612$ vs. $0.181$, median rank $267 \rightarrow 3$), achieves a $99.1\%$ physics-stability rate under Habitat-Sim with $\sim 20\times$ wall-time speed-up, and produces zero-overlap scenes when used as a drop-in asset bank for Holodeck. Controlled ablations on the same asset bank attribute the retrieval gain to description richness.

We present SS3D, a web-scale SfM-based self-supervision pretraining pipeline for feed-forward 3D estimation from monocular video. Our model jointly predicts depth, ego-motion, and intrinsics in a single forward pass and is trained/evaluated as a coherent end-to-end 3D estimator. To stabilize joint learning, we use an intrinsics-first two-stage schedule and a unified single-checkpoint evaluation protocol. Scaling SfM self-supervision to unconstrained web video is challenging due to weak multi-view observability and strong corpus heterogeneity; we address these with a multi-view signal proxy (MVS) used for filtering and curriculum sampling, and with expert training distilled into a single student. Pretraining on YouTube-8M (~100M frames after filtering) yields strong cross-domain zero-shot transfer and improved fine-tuning performance over prior self-supervised baselines. We release the pretrained checkpoint and code.

Test-time adaptation (TTA) has emerged as a promising paradigm for vision-language models (VLMs) to bridge the distribution gap between pre-training and test data. Recent works have focused on backpropagation-free TTA methods that rely on cache-based designs, but these introduce two key limitations. First, inference latency increases as the cache grows with the number of classes, leading to inefficiencies in large-scale settings. Second, suboptimal performance occurs when the cache contains insufficient or incorrect samples. In this paper, we present Prototype-Based Test-Time Adaptation (PTA), an efficient and effective TTA paradigm that uses a set of class-specific knowledge prototypes to accumulate knowledge from test samples. Particularly, knowledge prototypes are adaptively weighted based on the zero-shot class confidence of each test sample, incorporating the sample's visual features into the corresponding class-specific prototype. It is worth highlighting that the knowledge from past test samples is integrated and utilized solely in the prototypes, eliminating the overhead of cache population and retrieval that hinders the efficiency of existing TTA methods. This endows PTA with extremely high efficiency while achieving state-of-the-art performance on 15 image recognition benchmarks and 4 robust point cloud analysis benchmarks. For example, PTA improves CLIP's accuracy from 65.64% to 69.38% on 10 cross-domain benchmarks, while retaining 92% of CLIP's inference speed on large-scale ImageNet-1K. In contrast, the cache-based TDA achieves a lower accuracy of 67.97% and operates at only 50% of CLIP's inference speed.

Multimodal large language models (MLLMs) have advanced clinical tasks for common conditions, but their performance on rare diseases remains largely untested. In rare-disease scenarios, clinicians often lack prior clinical knowledge, forcing them to rely strictly on case-level evidence for clinical judgments. Existing benchmarks predominantly evaluate common-condition, single-image settings, leaving multimodal and multi-image evidence integration under rare-disease data scarcity systematically unevaluated. We introduce MMRareBench, to our knowledge the first rare-disease benchmark jointly evaluating multimodal and multi-image clinical capability across four workflow-aligned tracks: diagnosis, treatment planning, cross-image evidence alignment, and examination suggestion. The benchmark comprises 1,756 question-answer pairs with 7,958 associated medical images curated from PMC case reports, with Orphanet-anchored ontology alignment, track-specific leakage control, evidence-grounded annotations, and a two-level evaluation protocol. A systematic evaluation of 23 MLLMs reveals fragmented capability profiles and universally low treatment-planning performance, with medical-domain models trailing general-purpose MLLMs substantially on multi-image tracks despite competitive diagnostic scores. These patterns are consistent with a capacity dilution effect: medical fine-tuning can narrow the diagnostic gap but may erode the compositional multi-image capability that rare-disease evidence integration demands.

This paper presents an overview of the NTIRE 2026 Second Challenge on Day and Night Raindrop Removal for Dual-Focused Images. Building upon the success of the first edition, this challenge attracted a wide range of impressive solutions, all developed and evaluated on our real-world Raindrop Clarity dataset~\cite{jin2024raindrop}. For this edition, we adjust the dataset with 14,139 images for training, 407 images for validation, and 593 images for testing. The primary goal of this challenge is to establish a strong and practical benchmark for the removal of raindrops under various illumination and focus conditions. In total, 168 teams have registered for the competition, and 17 teams submitted valid final solutions and fact sheets for the testing phase. The submitted methods achieved strong performance on the Raindrop Clarity dataset, demonstrating the growing progress in this challenging task.

Open-vocabulary object detection (OVOD) enables models to recognize objects beyond predefined categories, but existing approaches remain limited in practical deployment. On the one hand, multimodal designs often incur substantial computational overhead due to their reliance on text encoders at inference time. On the other hand, tightly coupled training objectives introduce a trade-off between closed-set detection accuracy and open-world generalization. Thus, we propose Decoupled Cognition DETR (DeCo-DETR), a vision-centric framework that addresses these challenges through a unified decoupling paradigm. Instead of depending on online text encoding, DeCo-DETR constructs a hierarchical semantic prototype space from region-level descriptions generated by pre-trained LVLMs and aligned via CLIP, enabling efficient and reusable semantic representation. Building upon this representation, the framework further disentangles semantic reasoning from localization through a decoupled training strategy, which separates alignment and detection into parallel optimization streams. Extensive experiments on standard OVOD benchmarks demonstrate that DeCo-DETR achieves competitive zero-shot detection performance while significantly improving inference efficiency. These results highlight the effectiveness of decoupling semantic cognition from detection, offering a practical direction for scalable OVOD systems.

Medical imaging benchmarks often evaluate VLMs on pre-selected 2D images, slices, crops, or patches, making evaluation closer to visual recognition. Real clinical workflows impose a different burden: readers must search through complete studies, operate imaging software, navigate across slices and magnifications, and document visual evidence that can be audited. We argue that this evidence-producing workflow is a critical missing evaluation axis for medical imaging agents. To study it, we introduce MedFlowBench, a full-study benchmark for VLM agents, together with MedOpenClaw, a controlled and replayable runtime in which agents operate medical imaging viewers such as 3D Slicer and QuPath. In each episode, an agent inspects a complete radiology study or whole-slide pathology image, returns a task answer, and submits structured evidence, including key slices, coordinates, regions of interest, or lesion-state fields. This evidence is automatically checked against withheld masks, annotations, and labels. Across evaluated models, final answer-only scoring gives an overly optimistic picture: when answers must also be supported by correct evidence, performance drops substantially on complex workflows. We further find that adding image-analysis tools does not by itself solve the problem. Tools help when they make a complex procedure simple and reliable, but agents still struggle when they must choose inputs, manage viewer state, and verify intermediate outputs over multiple steps. MedFlowBench exposes whether medical imaging agents can produce auditable evidence from complete studies, rather than plausible answers from selected images.

Feature attributions often hide a critical modeling choice: they explain a prediction along a counterfactual path from a reference state to an input. Different baselines, interpolations, and generative trajectories define different paths and can therefor produce different explanations. We study this path ambiguity as a modeling problem. Our central question is whether the path can be chosen by the data-generating transport process, rather than by a hand-designed interpolation or by the sensitivity geometry of the model being explained. We separate attribution into fixed-path credit allocation and path selection. For a fixed path, we prove that the Aumann-Shapley line integral is the unique attribution rule under standard fixed-path axioms and explicit coordinate-trace regularity. For path selection, we minimize kinetic action over flows that transport a reference distribution to the data distribution, yielding a transport-geodesic attribution principle. We approximate this ideal with Rectified Flow and Reflow and derive stability bounds linking vector-field error to attribution error. Experiments show that lower-action, transport-consistent paths produce more stable and structured explanations, preserving competitive deletion faithfulness, without claiming data-manifold membership. Our code is available at https://github.com/cenweizhang/OTFlowSHAP.

Video reasoning requires a fine-grained understanding of the temporal dependencies and event-level relations between objects and events in videos. Current Multimodal Large Language Models (MLLMs) are prone to severe temporal hallucinations in video reasoning. An underlying cause of these hallucinations is weak visual-temporal grounding and the lack of explicit structure for modelling event relations. Models often rely on auxiliary text, such as dense captions, rather than explicitly anchoring their reasoning in actual visual evidence. However, these textual representations are inherently unstructured and fail to provide explicit causal constraints needed to guide the model's reasoning. In this work, we propose GraphThinker, a reinforcement finetuning method that constructs a structured event representation of a video and enforces visual grounding to jointly reduce reasoning hallucinations. Specifically, we employ an MLLM to construct an Event-based Video Scene Graph (EVSG) that captures both intra- and inter-event relations, guiding a structured video reasoning process. Moreover, we address the weak grounding issue by introducing a novel visual attention reward during reinforcement finetuning that encourages the model to actively attend to reliable visual cues. On the RexTime dataset, GraphThinker achieves an over 4% improvement in IoU=0.3 for moment localisation. On the VidHalluc dataset, GraphThinker achieves a 9.8% improvement in reducing temporal sequence hallucination and a 7.6% gain in Binary QA in reducing action hallucination, compared to the state-of-the-art methods.

Online Reinforcement Learning (RL) offers a promising avenue for complex image editing but is currently constrained by the scarcity of reliable and fine-grained reward signals. Existing evaluators frequently struggle with a critical perception gap we term "Attention Collapse," where models neglect cross-image comparisons and fail to capture fine-grained details, resulting in inaccurate perception and miscalibrated scores. To address these limitations, we propose SpatialReward, a reward model that enforces precise verification via explicit spatial reasoning. By anchoring reasoning to predicted edit regions, SpatialReward grounds semantic judgments in pixel-level evidence, significantly enhancing evaluative accuracy. Trained on a curated 260k spatial-aware dataset, our model achieves state-of-the-art performance on MMRB2 and EditReward-Bench, and outperforms proprietary evaluators on our proposed MultiEditReward-Bench. Furthermore, SpatialReward serves as a robust signal in online RL, boosting OmniGen2 by +0.90 on GEdit-Bench--surpassing the leading discriminative model and doubling the gain of GPT-4.1 (+0.45). These results demonstrate that spatial reasoning is essential for unlocking effective alignment in image editing.

This work introduces the first closed-loop adaptive optics (AO) system capable of optically correcting aberrations in real-time without a guidestar or a wavefront sensor. Nearly 40 years ago, Cederquist et al. demonstrated that asymmetric apertures enable phase retrieval (PR) algorithms to perform fully computational wavefront sensing, albeit at a high computational cost. More recently, Chimitt et al. extended this approach with machine learning and demonstrated real-time wavefront sensing using only a single (guidestar-based) point-spread-function (PSF) measurement. Inspired by these works, we introduce a guidestar-free AO framework built around asymmetric apertures and machine learning. Our approach combines three key elements: (1) an asymmetric aperture placed at the system's pupil plane that enables PR-based wavefront sensing, (2) a pair of machine learning algorithms that estimate the PSF from natural scene measurements and reconstruct phase aberrations, and (3) a spatial light modulator that performs optical correction. We experimentally validate this framework on dense natural scenes imaged through unknown obscurants. Our method outperforms state-of-the-art guidestar-free wavefront shaping methods, using an order of magnitude fewer measurements and three orders of magnitude less computation.

Lung cancer remains the leading cause of cancer mortality, driving the development of automated screening tools to alleviate radiologist workload. Standing at the frontier of this effort is Sybil, a deep learning model capable of predicting future risk solely from computed tomography (CT) with high precision. However, despite extensive clinical validation, current assessments rely purely on observational metrics. This correlation-based approach overlooks the model's actual reasoning mechanism, necessitating a shift to causal verification to ensure robust decision-making before clinical deployment. We propose S(H)NAP, a model-agnostic auditing framework that constructs generative interventional attributions validated by expert radiologists. By leveraging realistic 3D diffusion bridge modeling to systematically modify anatomical features, our approach isolates object-specific causal contributions to the risk score. Providing the first interventional audit of Sybil, we demonstrate that while the model often exhibits behavior akin to an expert radiologist, differentiating malignant pulmonary nodules from benign ones, it suffers from critical failure modes, including dangerous sensitivity to clinically unjustified artifacts and a distinct radial bias.

Anomaly detection usually assumes that abnormality is an intrinsic property of an observation. A defect is a defect, and a rare object is rare, regardless of where it appears. Many real-world anomalies do not work this way. A runner on a track is normal, but the same runner on a highway is not. The subject is unchanged; only the context makes it anomalous. This setting, long recognized as contextual anomaly detection, remains largely underexplored in modern vision-language systems. The difficulty is not merely empirical; it is formal. When anomaly labels depend on the relation between a subject and its context, any detector reasoning from a global representation that conflates subject and context is provably non-identifiable: two different subject-context configurations can map to the same embedding while requiring opposite labels, and no such detector can be correct on both. This impossibility motivates a different formulation: instead of asking whether an observation deviates from a global notion of normality, the model should ask whether subjects are compatible with their surrounding context. We define this as conditional compatibility learning. We instantiate this framework in CC-CLIP, a vision-language architecture that learns disentangled subject- and context-aware representations from a single image and fuses visual evidence through text-conditioned attention. CC-CLIP achieves state-of-the-art results on real-world contextual anomaly detection, substantially outperforming all existing CLIP-based and context-reasoning baselines. A single-branch variant of CC-CLIP also achieves competitive performance on structural anomaly benchmarks.

Posing 3D characters is a fundamental task in computer graphics. However, existing paradigms, ranging from traditional auto-rigging to recent pose-conditioned generative models, frequently struggle with inaccurate skinning weights, fixed mesh topologies, and poor pose conformance. These challenges have become particularly pronounced with the recent explosion of AI-generated 3D assets, which often exhibit flawed structures and fused geometry. To address these issues, we introduce Make-It-Poseable, a novel feed-forward framework that reformulates character posing as a skinning-free latent-space transformation problem. By decoupling shape deformation from the constraints of fixed mesh connectivity, our method directly operates on compact latent representations to reconstruct characters in target poses. To achieve this, our framework integrates a latent posing transformer for shape manipulation, a dense pose representation for fine-grained control, and an adaptive completion module optimized via a bipartite-matched latent loss to robustly handle topological changes. Extensive experiments demonstrate that our method significantly outperforms existing baselines in posing quality. Furthermore, our skeleton-agnostic design exhibits remarkable zero-shot generalization to diverse morphologies including quadrupeds and seamlessly supports various 3D authoring applications such as part replacement and refinement.

Accurate 6D pose estimation and tracking are core capabilities for physical AI systems, yet real-world deployment remains brittle and labor-intensive. Many pipelines rely on CAD models, manual masking, or per-object adaptation, and still fail under occlusion or fast motion without a principled way to recognize failure. We propose STORM, a unified framework for reference-conditioned 6D tracking that can operate from a single reference image, with minimal manual input and improved robustness. STORM combines: (i) Hierarchical Spatial Fusion Attention (HSFA), a task-driven reference-query fusion architecture that supports both single-reference and multi-reference conditioning and can optionally use vision-language semantic conditioning to resolve instance ambiguities; and (ii) a BCE-trained tracking verifier whose continuous compatibility logit is used as an energy-like score to detect drift and trigger automatic re-initialization. Experiments on LM-O and YCB-Video show that STORM improves annotation-free pose tracking accuracy over strong baselines and recovers reliably from severe occlusions and rapid viewpoint changes with minimal overhead.

Three-dimensional scene reconstruction from sparse-view satellite images is a long-standing and challenging task. While 3D Gaussian Splatting (3DGS) and its variants have recently attracted attention for its high efficiency, existing methods remain unsuitable for satellite images due to incompatibility with rational polynomial coefficient (RPC) models and limited generalization capability. Recent advances in generalizable 3DGS approaches show potential, but they perform poorly on multi-temporal sparse satellite images due to limited geometric constraints, transient objects, and radiometric inconsistencies. To address these limitations, we propose SkySplat, a novel self-supervised framework that integrates the RPC model into the generalizable 3DGS pipeline, enabling more effective use of sparse geometric cues for improved reconstruction. SkySplat relies only on RGB images and radiometric-robust relative height supervision, thereby eliminating the need for ground-truth height maps. Key components include a Cross-Self Consistency Module (CSCM), which mitigates transient object interference via consistency-based masking, and a multi-view consistency aggregation strategy that refines reconstruction results. Compared to per-scene optimization methods, SkySplat achieves an 86 times speedup over EOGS with higher accuracy. It also outperforms generalizable 3DGS baselines, reducing MAE from 13.18 m to 1.80 m on the DFC19 dataset significantly, and demonstrates strong cross-dataset generalization on the MVS3D benchmark. The is available at https://github.com/NanCheng2001/SkySplat-main

Novel view synthesis for underwater scene reconstruction presents unique challenges due to complex light-media interactions. Optical scattering and absorption in water body bring inhomogeneous medium attenuation interference that disrupts conventional volume rendering assumptions of uniform propagation medium. While 3D Gaussian Splatting (3DGS) offers real-time rendering capabilities, it struggles with underwater inhomogeneous environments where scattering media introduces artifacts and inconsistent appearance. In this study, we propose a physics-based framework that disentangles object appearance from water medium effects through tailored Gaussian modeling. Our approach introduces appearance embeddings, which are explicit medium representations for backscatter and attenuation, enhancing scene consistency. In addition, we propose a depth-guided optimization strategy that leverages pseudo-depth maps as supervision with depth regularization and scale penalty terms to improve geometric fidelity. By integrating the proposed appearance and medium modeling components via an underwater imaging model, our approach achieves both high-quality novel view synthesis and physically accurate scene restoration. Experiments demonstrate our significant improvements in rendering quality and restoration accuracy over existing methods. The project page is available at https://bilityniu.github.io/3D-UIR.

Multimodal Large Language Models (MLLMs) have achieved significant advances in integrating visual and linguistic information, yet their ability to reason about complex and real-world scenarios remains limited. The existing benchmarks are usually constructed in the task-oriented manner without guarantee that different task samples come from the same data distribution, thus they often fall short in evaluating the synergistic effects of lower-level perceptual capabilities on higher-order reasoning. To lift this limitation, we contribute Lens, a multi-level benchmark with 3.4K contemporary images and 60K+ human-authored questions covering eight tasks and 12 daily scenarios, forming three progressive task tiers, i.e., perception, understanding, and reasoning. One feature is that each image is equipped with rich annotations for all tasks. Thus, this dataset intrinsically supports to evaluate MLLMs to handle image-invariable prompts, from basic perception to compositional reasoning. In addition, our images are manully collected from the social media, in which 53% were published later than Jan. 2025. We evaluate 15+ frontier MLLMs such as Qwen2.5-VL-72B, InternVL3-78B, GPT-4o and two reasoning models QVQ-72B-preview and Kimi-VL. These models are released later than Dec. 2024, and none of them achieve an accuracy greater than 60% in the reasoning tasks. Project page: https://github.com/Lens4MLLMs/lens. ICCV 2025 workshop page: https://lens4mllms.github.io/mars2-workshop-iccv2025/

Deep neural networks have recently been used to edit images with great success, in particular for faces. However, they are often limited to only being able to work at a restricted range of resolutions. Many methods are so flexible that face edits can often result in an unwanted loss of identity. This work proposes to learn how to perform semantic image edits through the application of smooth warp fields. Previous approaches that attempted to use warping for semantic edits required paired data, i.e. example images of the same subject with different semantic attributes. In contrast, we employ recent advances in Generative Adversarial Networks that allow our model to be trained with unpaired data. We demonstrate face editing at very high resolutions (4k images) with a single forward pass of a deep network at a lower resolution. We also show that our edits are substantially better at preserving the subject's identity. The robustness of our approach is demonstrated by showing plausible image editing results on the Cub200 birds dataset. To our knowledge this has not been previously accomplished, due the challenging nature of the dataset.

Fire effects are widely used in various computer graphics applications such as visual effects and video games. Modeling the shape and appearance of fire phenomenon is challenging as the underlying effects are driven by complex laws of physics. State-of-the-art fire modeling techniques rely on sophisticated physical simulations which require intensive parameter tuning, or use simplifications which produce physically invalid results. In this paper, we present a novel method of reconstructing physically valid fire models from multi-view stereo images. Our method, for the first time, provides plausible estimation of physical properties (e.g., temperature, density) of a fire volume using RGB cameras. This allows for a number of novel phenomena such as global fire illumination effects. The effectiveness and usefulness of our method are tested by generating fire models from a variety of input data, and applying the reconstructed fire models for realistic illumination of virtual scenes.

Variational auto-encoders (VAEs) are a popular and powerful deep generative model. Previous works on VAEs have assumed a factorized likelihood model, whereby the output uncertainty of each pixel is assumed to be independent. This approximation is clearly limited as demonstrated by observing a residual image from a VAE reconstruction, which often possess a high level of structure. This paper demonstrates a novel scheme to incorporate a structured Gaussian likelihood prediction network within the VAE that allows the residual correlations to be modeled. Our novel architecture, with minimal increase in complexity, incorporates the covariance matrix prediction within the VAE. We also propose a new mechanism for allowing structured uncertainty on color images. Furthermore, we provide a scheme for effectively training this model, and include some suggestions for improving performance in terms of efficiency or modeling longer range correlations.

Embodied agents in household environments must plan under partial observation: they need to remember objects, track state changes, and recover when actions fail. Existing benchmarks only partially test this ability. Egocentric video datasets capture realistic human activities but remain passive, while interactive simulators support execution but rely on synthetic scenes and hand-crafted dynamics, introducing a sim-to-real gap and often assuming fully observable state. We introduce Ego2World, an executable benchmark that turns egocentric cooking videos into executable symbolic worlds governed by graph-transition rules. Built on HD-EPIC, Ego2World derives reusable transition rules from video annotations and executes them in a hidden symbolic world graph. During evaluation, the simulator maintains the hidden world graph, while the agent plans over its own partial belief graph using only local observations and execution feedback. This separation forces agents to update memory and replan without observing the true world state. Experiments show that action-overlap scores overestimate physical-state success, and that persistent belief memory improves task completion while reducing repeated visual exploration -- suggesting that belief maintenance should be a first-class target of embodied-agent evaluation.

Text-driven motion diffusion models are capable of generating realistic human motions, but text alone often struggles to express fine-level nuances of motion, commonly referred to as style. Recent approaches have tackled this challenge by attaching a style injection mechanism to a pretrained text-driven diffusion model. Existing stylization methods, however, either require style-specific fine-tuning of existing models or rely on heavy ControlNet-based architectures, limiting efficiency and generalization to unseen styles. We propose a lightweight style conditioning framework that dynamically modulates a pretrained diffusion model through hypernetwork-generated LoRA parameters. A style reference motion is encoded into a global style embedding, which is mapped by a hypernetwork to low-rank updates applied at each denoising step of the diffusion model. By structuring the style latent space with a supervised contrastive loss, our framework reliably captures diverse stylistic attributes, improves generalization to unseen styles, and supports optimization-based guidance without requiring predefined style categories. Experiments on the HumanML3D and 100STYLE datasets show state-of-the-art stylization results, while achieving improved stylization for unseen styles.

Vision-and-Language Navigation (VLN) is a cornerstone of embodied intelligence. However, current agents often suffer from significant performance degradation when transitioning from simulation to real-world deployment, primarily due to perceptual instability (e.g., lighting variations and motion blur) and under-specified instructions. While existing methods attempt to bridge this gap by scaling up model size and training data, we argue that the bottleneck lies in the lack of robust spatial grounding and cross-domain priors. In this paper, we propose StereoNav, a robust Vision-Language-Action framework designed to enhance real-world navigation consistency. To address the inherent gap between synthetic training and physical execution, we introduce Target-Location Priors as a persistent bridge. These priors provide stable visual guidance that remains invariant across domains, effectively grounding the agent even when instructions are vague. Furthermore, to mitigate visual disturbances like motion blur and illumination shifts, StereoNav leverages stereo vision to construct a unified representation of semantics and geometry, enabling precise action prediction through enhanced depth awareness. Extensive experiments on R2R-CE and RxR-CE demonstrate that StereoNav achieves state-of-the-art egocentric RGB performance, with SR and SPL scores of 81.1% and 68.3%, and 67.5% and 52.0%, respectively, while using significantly fewer parameters and less training data than prior scaling-based approaches. More importantly, real-world robotic deployments confirm that StereoNav substantially improves navigation reliability in complex, unstructured environments. Project page: https://yunheng-wang.github.io/stereonav-public.github.io.

Action diffusion excels at high-fidelity action generation but incurs heavy computational costs owing to its iterative denoising nature. Despite current technologies showing promise in accelerating diffusion transformers by reusing the cached features, they struggle to adapt to policy dynamics arising from diverse perceptions and multi-round rollout iterations in open environments. We propose test-time sparsity to tackle this challenge, which aims to accelerate action diffusion by dynamically predicting prunable residual computations for each model forward at test time. However, two bottlenecks remain in this paradigm: 1) repetitive conditional encoding and pruning offset most potential speed gains, and 2) the features cached from previous denoising timesteps cannot constrain large pruning errors under aggressive sparsity. To address the first bottleneck, we design a highly parallelized inference pipeline that minimizes the non-decoder delay to milliseconds. Specifically, we first design a lightweight pruner that shares the encoder with the diffusion transformer. Then, we decouple the encoding and pruning from the autoregressive denoising loop by processing all denoising timesteps in parallel, and overlap the pruner with the decoder forward inference through asynchronism. To overcome the second bottleneck, we introduce an omnidirectional reusing strategy, which achieves 95% sparsity by selectively reusing features cached from the current forward, previous denoising timesteps, and earlier rollout iterations. To learn the rollout-level reusing strategies, we sample a few action trajectories to supervise the sparsified diffusion step by step. Extensive experiments demonstrate that our method reduces FLOPs by 92% and accelerates action generation by 5x, achieving lossless performance with an inference frequency of 47.5 Hz. Our code is available at https://github.com/ky-ji/Test-time-Sparsity.

Illuminant estimation aims to infer scene illumination from image measurements despite intrinsic ambiguities between surface reflectance and lighting. Most existing methods operate on trichromatic RGB images and are therefore fundamentally limited by the restricted spectral information available. Hyperspectral imaging provides a much richer representation of scene radiance and has the potential to alleviate these ambiguities. However, its high dimensionality poses computational and statistical challenges. In this work, we systematically study the effect of spectral dimensionality and representation choice on illuminant estimation performance using hyperspectral data. We adopt the practical and effective Color-by-Correlation (CbC) framework as the estimation backbone and analyze its behavior under different spectral dimensionality reduction strategies. Our results offer practical insights into how hyperspectral information can be efficiently exploited for illuminant estimation and identify conditions under which compact spectral representations outperform conventional RGB-based approaches. The code is available at https://github.com/IVRL/Reduced-Spectral-Color-Constancy.

Boundary Representation (BRep) is the standard format for Computer-Aided Design (CAD), yet reconstructing high-quality BReps from single-view images remains challenging due to the complexity of topological constraints and operation sequences. We present Img2CADSeq, a multi-stage pipeline that overcomes these limitations by encoding CAD sequences into a three-level hierarchical codebook. Guided by an importance prioritization, this strategy values profiles over details, compressing long sequences into a stable discrete latent space. To bridge the modality gap, we leverage a coarse-to-fine point cloud intermediate, aligning 2D visual features with 3D CAD sequences via contrastive learning to condition a VQ-Diffusion model. Supported by newly introduced CAD-220K and PrintCAD datasets, our approach ensures robust industrial domain adaptation. Extensive experiments demonstrate that Img2CADSeq significantly outperforms state-of-the-art methods, producing standard STEP files that can be directly used in commercial CAD software.

Visual evidence selection is a critical component of multimodal retrieval-augmented generation (RAG), yet existing methods typically rely on semantic relevance or surface-level similarity, which are often misaligned with the actual utility of visual evidence for downstream reasoning. We reformulate multimodal evidence selection from an information-theoretic perspective by defining evidence utility as the information gain induced on a model's output distribution. To overcome the intractability of answer-space optimization, we introduce a latent notion of evidence helpfulness and theoretically show that, under mild assumptions, ranking evidence by information gain on this latent variable is equivalent to answer-space utility. We further propose a training-free, surrogate-accelerated framework that efficiently estimates evidence utility using lightweight multimodal models. Experiments on MRAG-Bench and Visual-RAG across multiple model families demonstrate that our method consistently outperforms state-of-the-art RAG baselines while achieving substantial reductions in computational cost.

Video understanding requires active evidence seeking, motivating tool-augmented video agents for temporal reasoning, cross-modal understanding, and complex question answering. Existing video agents have improved video reasoning with retrieval, memory, frame inspection, and verifier tools, but they still face two limitations: (1) a coarse tool space that lacks fine-grained operations for compositional reasoning; and (2) a flat action space that forces high-level video intents into primitive executable tool calls. In this paper, we address these challenges with two complementary designs. First, we construct a MetaAug-Video Tool Library (MVTL), an extensible tool library with 134 registered tools, including 26 base tools for general multimodal signal processing and 108 meta tools for filtering, aggregation, reranking, formatting, and other intermediate-result operations. MVTL supports dual-level access to both structured video information and raw modal evidence, enabling diverse video reasoning scenarios. Second, we propose ReTool-Video, a recursive tool-using method that grounds high-level video intents into executable tool chains. In ReTool-Video, matched actions are executed directly, while unmatched intents are delegated to a resolver for parameter repair, tool substitution, or decomposition. This allows abstract actions such as temporal merging, cross-modal verification, or repeated-event aggregation to be progressively translated into concrete multimodal operations at runtime. Experiments on MVBench, MLVU, and Video-MME w/o sub. show that ReTool-Video consistently outperforms strong baselines. Further analysis demonstrates that recursive grounding and fine-grained meta tools improve the stability and effectiveness of complex video understanding.

Text-to-image (T2I) generation has advanced rapidly, making reliable evaluation critical as performance differences between models narrow. Existing evaluation practices typically apply uniform annotation mechanisms, such as Likert-scale or binary question answering (BQA), across heterogeneous evaluation skills, despite fundamental differences in their nature. In this work, we revisit T2I evaluation through the lens of skill-aligned annotation, where annotation strategies reflect the underlying characteristics of each evaluation skill. We systematically compare skill-aligned annotation against uniform baselines and show that it produces more consistent evaluation signals, with higher inter-annotator agreement and improved stability across models. Finally, we present an automated pipeline that instantiates the proposed evaluation protocol, enabling scalable and fine-grained evaluation with spatially grounded feedback. Our work highlights that improving the foundations of image evaluation can increase reliability and efficiency without simply scaling annotation effort. We hope this motivates further research on refining evaluation protocols as a central component of reliable model assessment.

Few-shot action recognition (FSAR) requires models to generalize to novel action categories from only a handful of annotated samples. Despite progress with vision-language models, existing approaches still suffer from semantic-temporal misalignment, where static textual prompts fail to capture decisive visual cues that appear sparsely across sequences, and from inadequate modeling of multi-scale temporal dynamics, as short-term discriminative cues and long-range dependencies are often either oversmoothed or fragmented. To address these challenges, we propose Semantic Temporal Adaptive Representation Learning (STAR), a unified framework, consisting of a semantic-alignment component and a temporal-aware component, effectively bridging the semantic and temporal gaps and transferring the sequence modeling capability of Mamba into the FSAR. The semantic alignment module introduces a Temporal Semantic Attention (TSA) mechanism, which performs frame-level cross-modal alignment with textual cues, ensuring fine-grained semantic-temporal consistency. The temporal-aware module incorporates a Semantic Temporal Prototype Refiner (STPR) that integrates semantic-guided Mamba blocks with multi-frequency temporal sampling and bidirectional state-space refinement, yielding semantically aligned prototypes with enhanced discriminative fidelity and temporal consistency. Furthermore, temporally dependent class descriptors derived from large language models (LLMs) provide long-range semantic guidance. Extensive experiments on five FSAR benchmarks demonstrate the consistent superiority of STAR over state-of-the-art methods. For instance, STAR achieves up to 8.1% and 6.7% gains on the SSv2-Full and SSv2-Small datasets under the 1-shot setting, and 7.3% on HMDB51, validating its effectiveness under limited supervision. The code is available at https://github.com/HongliLiu1/STAR-main.

Diffusion-based models have shown strong performance in video super-resolution (VSR) and video frame interpolation (VFI). However, their role in the coupled space-time video super-resolution (STVSR) setting remains limited. Existing diffusion-based STVSR approaches suffer from two issues: (1) low inference efficiency and (2) insufficient utilization of spatiotemporal information. These limitations impede deployment. To address these issues, we introduce DiffST, an efficient spatiotemporal-aware video diffusion framework for real-world STVSR. To improve efficiency, we adapt a pre-trained diffusion model for one-step sampling and process the entire video directly rather than operating on individual frames. Furthermore, to enhance spatiotemporal information utilization, we introduce cross-frame context aggregation (CFCA) and video representation guidance (VRG). The CFCA module aggregates information across multiple keyframes to produce intermediate frames. The VRG module extracts video-level global features to guide the diffusion process. Extensive experiments show that DiffST obtains leading results on real-world STVSR tasks. It also maintains high inference efficiency, running about 17$\times$ faster than previous diffusion-based STVSR methods. Code is available at: https://github.com/zhengchen1999/DiffST.

The Engram module -- a hash-keyed, O(1) associative memory injected into Transformer layers -- was recently shown to improve large language model pretraining, with the appealing interpretation that it provides a content-addressed shortcut to recurring local token patterns. We ask whether this interpretation transfers to autoregressive (AR) image generation, or whether the observed gains, if any, come from a different mechanism. We adapt the Engram module to vision with 2D spatial $n$-gram hashing, gated fusion, and KV-cache-compatible incremental inference, and inject it into a class-conditional AR generator trained on ImageNet 256x256. Across a sweep of backbone-to-memory budget ratios $ρ{\in}[0.17, 0.90]$, every Engram-augmented variant trails the pure AR baseline in FID, indicating that the module saves backbone FLOPs but does not, by itself, improve sample quality. We then probe how the module is used. A gate-clamp sweep shows that disabling the Engram pathway entirely is catastrophic, yet a tiny constant gate (g=0.10) matches or beats the learned gate -- inconsistent with a heavily content-addressed recall mechanism. A donor-probe experiment shows that swapping the hash inputs for matched, adversarial, or random same-class exemplars produces statistically indistinguishable next-token distributions, while collapsing or randomising the table degrades them by two to three orders of magnitude. Finally, training a model from scratch with the entire memory table frozen to $\mathcal{N}(0, 1)$ noise costs only $Δ\text{FID}{=}0.10$ and actually raises Inception Score. Together, these findings indicate that the Engram in AR image generation behaves not as a content-addressed retriever but as a gated architectural side-pathway: a hash-keyed residual stream whose benefit is dominated by the pathway itself, with the learned table contributing only a small distributional refinement.

Foundation models and low-rank adapters enable efficient on-device generative AI but raise risks such as intellectual property leakage and model recovery attacks. Existing defenses are often impractical because they require retraining or access to the original dataset. We propose LoREnc, a training-free framework that secures both FMs and adapters via spectral truncation and compensation. LoREnc suppresses dominant low-rank components of FM weights, compensates for the missing information in authorized adapters, and further applies orthogonal reparameterization to obscure structural fingerprints of the protected adapter. Unauthorized users produce structurally collapsed outputs, while authorized users recover exact performance. Experiments demonstrate that LoREnc provides strong protection against model recovery with under 1% computational overhead.

Image restoration under multiple adverse weather conditions aims to develop a single model to recover the underlying scene with high visibility. Weather-related artifacts vary with the particle's distance to the camera according to the established scene visibility analysis, where close and faraway regions are more affected by falling drops and fog effects, respectively. Existing methods fail to consider this weather-specific physical visual process; thus, the restoration performance is limited. In this work, we analyze the common visual factors in adverse weather conditions and present a unified imaging model that considers the individually visible particles and fog-like aggregate scattering effects. Further, we design a novel weather-prior-based network, which leverages the weather-related prior information to help recover the scene by enhancing the features using the estimated occlusion and transmission. Experimental results in multiple adverse scenarios show the superiority of our method against state-of-the-art methods.

Vision-language models (VLMs) have demonstrated remarkable capabilities in bridging visual perception and natural language understanding, enabling a wide range of multimodal reasoning tasks. However, they often produce object hallucinations, describing content absent from the input image, which limits their reliability and interpretability. To address this limitation, we propose Dual-Pathway Circuit Analysis, a framework that identifies and characterizes hallucination-related circuits in VLMs for mechanistic understanding and causal probing. We first apply activation patching across five architecturally diverse VLMs to identify a visual grounding pathway that supports correct predictions and a hallucination pathway that drives erroneous outputs. We then introduce Conditional Pathway Analysis (CPA) to characterize pathway-level interactions, revealing that grounding components remain strongly redundant in both correct and hallucinating samples but undergo a consistent polarity flip, shifting from supporting the ground truth on correct samples to aligning with the hallucinated answer on erroneous ones. We further perform targeted suppression of hallucination-pathway components, showing that scaling these components reduces object hallucination by up to 76% with minimal accuracy cost, and validate that the same circuit selectively transfers to relational but not attribute hallucination. Evaluations on POPE-adversarial and AMBER show that the identified circuits are consistent across architectures, support causal intervention, and transfer selectively across hallucination types.

Text-to-image generation models have achieved remarkable progress in preference optimization, yet achieving robust alignment across diverse reward models remains a significant challenge. Existing multi-reward fusion approaches rely on weighted summation, which is costly to tune and insufficient for balancing conflicting objectives. More critically, optimization with reward models is highly susceptible to reward hacking, where reward scores increase while the perceived quality of generated images deteriorates. We demonstrate that optimizing against a unified global target under heterogeneous reward upper bounds can induce reward hacking, a risk further exacerbated by the inherent instability of weak reward models. To mitigate this, we propose a Pareto Frontier-Guided Optimal Transport (PG-OT) framework. Our method constructs a prompt-specific Pareto frontier and maps dominated samples toward it via distribution-aware optimal transport. Furthermore, we develop both online and offline optimization strategies tailored to diverse reward signal characteristics. To provide a more rigorous assessment, we introduce the Joint Domination Rate (JDR) and Joint Collapse Rate (JCR) as principled metrics to quantify multi-reward synergy and reward hacking. Experimental results show that our approach outperforms strong baselines with an 11% gain in JDR and achieves a near 80% win rate in human evaluations.

We introduce EvObj for unsupervised 3D instance segmentation that bridges the geometric domain gap between synthetic pretraining data and real-world point clouds. Current methods suffer from structural discrepancies when transferring object priors from synthetic datasets (e.g., ShapeNet) to real scans (e.g., ScanNet), particularly due to morphological variations and occlusion artifacts. To address this, EvObj integrates two innovative modules: (1) An object discerning module that dynamically refines object candidates, enabling continuous adaptation of object priors to target domains; and (2) An object completion module that reconstructs partial geometries after discovering objects. We conduct extensive experiments on both real-world and synthetic datasets, demonstrating superior 3D object segmentation performance over all baselines while achieving state-of-the-art results.

Category-agnostic pose estimation (CAPE) aims to localize keypoints on query images from arbitrary categories, using only a few annotated support examples for guidance. Recent approaches either treat keypoints as isolated entities or rely on manually defined skeleton priors, which are costly to annotate and inherently inflexible across diverse categories. Such oversimplification limits the model's capacity to capture instance-wise structural cues critical for accurate pixel-level localization. To overcome these limitations, we propose GenCape, a Generative-based framework for CAPE that infers keypoint relationships solely from image-based support inputs, without additional textual descriptions or predefined skeletons. Our framework consists of two principal components: an iterative Structure-aware Variational Autoencoder (i-SVAE) and a Compositional Graph Transfer (CGT) module. The former infers soft, instance-specific adjacency matrices from support features through variational inference, embedded layer-wise into the Graph Transformer Decoder for progressive structural priors refinement. The latter adaptively aggregates multiple latent graphs into a query-aware structure via Bayesian fusion and attention-based reweighting, enhancing resilience to visual uncertainty and support-induced bias. This structure-aware design facilitates effective message propagation among keypoints and promotes semantic alignment across object categories with diverse keypoint topologies. Experimental results on the MP-100 dataset show that our method achieves substantial gains over graph-support baselines under both 1- and 5-shot settings, while maintaining competitive performance against text-support counterparts.

Understanding why deep neural networks (DNNs) fail to generalize to unseen samples remains a long-standing challenge. Existing studies mainly examine changes in externally observable factors such as data, representations, or outputs, yet offer limited insight into how a model's internal decision mechanism evolves from training to test. To address this gap, we introduce Decision Pattern Shift (DPS), a new perspective that defines generalization through the stability of internal decision patterns and quantifies failure as their deviation from those learned during training. Specifically, we represent each sample's decision pattern as a GradCAM-based channel-contribution vector, which captures how feature channels collectively support a prediction, and we propose the DPS metric to measure its discrepancy from the class-average pattern. Empirical analyses across multiple datasets and architectures show that, (i) decision patterns form a highly structured, class-consistent space with strong intra-class cohesion and low inter-class confusion, enabling direct analysis of a model's decision logic; (ii) the DPS magnitude correlates linearly with the generalization gap (nearly all Pearson r > 0.8), revealing generalization as a systematic drift in the model's internal decision mechanism; (iii) the DPS spectrum organizes diverse generalization degradation scenarios (covering ideal generalization, in-distribution degradation, domain shift, out-of-distribution, and shortcut learning) into a continuous trajectory, providing a unified explanation of their failure modes. These findings open up new possibilities for early generalization-risk detection, failure-mode diagnosis, and channel-level defect localization.

Artificial intelligence (AI) has transformed imaging inverse problems, from medical diagnostics to Earth observation. Yet deep neural networks can produce hallucinations, realistic-looking but incorrect details, undermining their reliability, especially when ground truth data is unavailable. We develop a theoretical framework showing that such hallucinations are not merely artifacts of particular models, but can arise from the ill-posed nature of the inverse problem itself. We derive necessary and sufficient conditions for hallucinations, together with computable bounds on their magnitude that depend only on the forward model. Building on this theory, we introduce algorithms to: (1) estimate the minimum hallucination magnitude achievable by any reconstruction model for a given input; (2) assess the faithfulness of reconstructed details by a given reconstruction model. Experiments across three imaging tasks demonstrate that our approach applies broadly, including to modern generative models, and provides a principled way to quantify and evaluate AI hallucinations.

General object detection (OD) struggles to detect objects in the target domain that differ from the training distribution. To address this, recent studies demonstrate that training from multiple source domains and explicitly processing them separately for multi-source domain adaptation (MSDA) outperforms blending them for unsupervised domain adaptation (UDA). However, existing MSDA methods learn domain-agnostic features from domain-specific RGB images while preserving domain-specific information from the domain-agnostic feature map. To address this, we propose MS-DePro: Multi-Source Detector with Depth and Prompt, composed of (1) depth-guided localization and (2) multi-modal guided prompt learning. We leverage domain-agnostic input modalities, namely depth maps and text, to encode domain-agnostic characteristics. Specifically, we utilize depth maps to generate domain-agnostic region proposals for localization and integrate multi-modal features to align learnable text embeddings for classification. MS-DePro achieves state-of-the-art performance on MSDA benchmarks, and comprehensive ablations demonstrate the effectiveness of our contributions. Our code is available on https://github.com/sejong-rcv/Multi-Modal-Guided-Multi-Source-Domain-Adaptation-for-Object-Detection.

Recent 3D generative models can synthesize high-quality assets, but their outputs are typically static: they lack the skeletal rigs, joint hierarchies, and skinning weights required for animation. This limits their use in games, film, simulation, virtual agents, and embodied AI, where assets must not only look plausible but also move plausibly. We introduce Rigel3D, a generative method for animation-ready 3D assets represented as rigged meshes. Unlike post-hoc auto-rigging methods that attach rigs to completed shapes, our method jointly models geometry and rig structure through coupled surface and skeleton structured latent representations. A rig-aware autoencoder decodes these representations into mesh geometry, skeleton topology, joint coordinates, and skinning weights, while a two-stage latent generative model synthesizes both surface and skeleton representations for image-conditioned generation. To support downstream animation workflows, we further introduce an open-vocabulary joint labeling module that embeds generated joints into a shared vision-language space, enabling correspondence to arbitrary retargeting templates. Experiments on large-scale rigged asset datasets demonstrate that our method generates diverse, high-quality animation-ready assets and outperforms existing rigging baselines across multiple metrics.

Instruction-based image editing (IIE) models have recently demonstrated strong capability in modifying specific image regions according to natural language instructions, which implicitly requires identifying where an edit should be applied. This indicates that such models inherently perform language-conditioned visual semantic grounding. In this work, we investigate whether this implicit grounding can be leveraged for zero-shot referring image segmentation (RIS), a task that requires pixel-level localization of objects described by natural language expressions. Through systematic analysis, we reveal that strong foreground-background separability emerges in the internal representations of these models at the earliest denoising timestep, well before any visible image transformation occurs. Building on this insight, we propose a training-free framework that repurposes pretrained image editing models for RIS by exploiting their intermediate representations. Our approach decomposes localization into two complementary components: attention-based spatial priors that estimate where to focus, and feature-based semantic discrimination that determines what to segment. By leveraging feature-space separability, the framework produces accurate segmentation masks using only a single denoising step, without requiring full image synthesis. Extensive experiments on RefCOCO, RefCOCO+, and RefCOCOg demonstrate that our method achieves superior performance over existing zero-shot baselines.

Vision-language-action (VLA) models are effective robot action executors, but they remain limited on long-horizon tasks due to the dual burden of extended closed-loop planning and diverse physical operations. We therefore propose VLAs-as-Tools, a strategy that distributes this burden across a high-level vision language model (VLM) agent for temporal reasoning and a family of specialized VLA tools for diverse local physical operations. The VLM handles scene analysis, global planning, and recovery, while each VLA tool executes a bounded subtask. To tightly couple agent planning with VLA tool execution in long-horizon tasks, we introduce a VLA tool-family interface that exposes explicit tool selection and in-execution progress feedback, enabling efficient event-triggered agent replanning without continuous agent polling. To obtain diverse specialized VLA tools that faithfully follow agent invocations, we further propose Tool-Aligned Post-Training (TAPT), which constructs invocation-aligned training units for instruction following and adopts tool-family residual adapters for efficient tool specialization. Experiments show that VLAs-as-Tools improves the success rate of $π_{0.5}$ by 4.8 points on LIBERO-Long and 23.1 points on RoboTwin, and further enhances invocation fidelity by 15.0 points as measured by Non-biased Rate. Code will be released.

Autoregressive video generation enables streaming and open-ended long video synthesis, but still suffers from long-term degradation caused by accumulated errors. Existing KVCache strategies usually apply unified historical-frame retention, implicitly assuming homogeneous historical dependencies across attention heads. We revisit historical-frame attention and reveal three distinct head types: Anchor Heads require broad long-range context, Wave Heads exhibit periodic temporal dependencies, and Veil Heads focus on initial and adjacent frames. Based on this finding, we propose Pyramid Forcing, a head-aware pyramidal KVCache framework that identifies head types offline, assigns behavior-specific cache policies, and supports heterogeneous cache lengths via efficient ragged-cache attention. Experiments on Self Forcing and Causal Forcing show that Pyramid Forcing consistently improves long-horizon generation quality on VBench-Long, increasing the 60-second Self Forcing score from 77.87 to 81.21 while enhancing motion dynamics, visual fidelity, and semantic consistency. Project: https://if-lab-pku.github.io/Pyramid-Forcing/.

Face presentation attack detection (FacePAD) remains challenging under diverse spoofing representation, including 2D print and replay, 3D mask-based spoofing, makeup-induced appearance manipulation, and physical occlusions, as well as under varying capture conditions. Motion cues are highly discriminative for FacePAD but typically require explicit optical flow estimation, which introduces substantial computational overhead and limits real-time deployment. In this work, we leverage optical flow to enhance motion representation during training while eliminating the need for flow computation at inference. We propose a dual-branch teacher model that fuses appearance cues from RGB frames with motion cues derived from colorwheel-encoded optical flow, enabling effective modeling of micro-motions and temporal consistency. To enable efficient deployment, we introduce a knowledge distillation framework that transfers motion-aware knowledge from the flow-augmented teacher to a lightweight RGB-only student via logit distillation. As a result, the student implicitly learns motion-sensitive representations without requiring explicit flow estimation or additional feature extraction blocks at inference. Extensive experiments demonstrate strong performance across multiple benchmarks, achieving 0.0% HTER on Replay-Attack and Replay-Mobile, 0.94% HTER on ROSE-Youtu, 5.65% HTER on SiW-Mv2, and 0.42% ACER on OULU-NPU. The distilled student achieves performance comparable to or better than the teacher while significantly reducing parameters and FLOPs, achieving 52 FPS on an NVIDIA Jetson Orin Nano, indicating its suitability for real-time and resource-constrained FacePAD deployment.

Generalizable 3D Gaussian Splatting has recently emerged as an efficient approach for novel-view synthesis, enabling feed-forward synthesis from only a few input views. However, existing pixel-wise feed-forward methods suffer from over-bright renderings when the number of input views varies during inference, as well as insufficient supervision for accurate Gaussian scale estimation, which leads to hole artifacts, particularly in high-resolution renderings. To address these issues, we identify that the over-brightness is caused by the varying number of overlapping Gaussians and propose a simple alpha normalization strategy to maintain brightness consistency across different number of input views. In addition, we introduce an auxiliary 3D sampling-based regularizer to improve Gaussian scale estimation, thereby mitigating hole artifacts in high-resolution rendering. Experiments on benchmark datasets demonstrate that our method significantly improves baseline models under varying input-view and high-resolution rendering settings.

When humans describe a visual scene, they do not process the entire image uniformly; instead, they selectively fixate on regions relevant to their intended description. In contrast, current multimodal large language models (MLLMs) attend to all visual tokens at each generation step, leading to diluted focus and unnecessary computational overhead. In this work, we introduce Gaze Attention, a novel mechanism that enables MLLMs to selectively attend to task-relevant visual regions during generation. Specifically, we spatially group visual embeddings-stored as key-value caches-into compact gaze regions, each represented by a lightweight descriptor. At each decoding step, the model dynamically selects the most relevant regions and restricts attention to them, reducing redundant computation while enhancing focus. To mitigate the loss of global context caused by localized attention, we further propose learnable context tokens appended to each image or frame, allowing the model to maintain holistic visual awareness. Extensive experiments on image and video understanding benchmarks demonstrate that Gaze Attention matches or surpasses dense-attention baselines, while using up to 90% fewer visual KV entries in the attention computation.

Recent image editing models have achieved remarkable progress in instruction following, multimodal understanding, and complex visual editing. However, existing benchmarks often fail to faithfully reflect human judgment, especially for strong frontier models, due to limited task difficulty and coarse-grained evaluation protocols. In parallel, reward models have become increasingly important for RL-based image editing optimization, yet existing reward model benchmarks still rely on unrealistic evaluation settings that deviate from practical RL scenarios. These limitations hinder reliable assessment of both image editing models and reward models. To address these challenges, we introduce Edit-Compass and EditReward-Compass, a unified evaluation suite for image editing and reward modeling. Edit-Compass contains 2,388 carefully annotated instances spanning six progressively challenging task categories, covering capabilities such as world knowledge reasoning, visual reasoning, and multi-image editing. Beyond broad task coverage, Edit-Compass adopts a fine-grained multidimensional evaluation framework based on structured reasoning and carefully designed scoring rubrics. In parallel, EditReward-Compass contains 2,251 preference pairs that simulate realistic reward modeling scenarios during RL optimization.

Clinical diagnostic workups typically follow a modality escalation pathway: after initial clinical evaluation, clinicians begin with routine structural imaging (e.g., MRI), selectively add sequences such as FLAIR or T2 to refine the differential, and reserve molecular imaging (e.g., amyloid-PET) for cases that remain uncertain after standard evaluation. Consequently, patients are observed with heterogeneous and often incomplete modality subsets. However, most current AI models assume fixed data modalities as the model inputs. In this paper, we present BrainAnytime, a unified pretraining framework pretrained on 34,899 3D brain scans from five datasets that support brain image analysis under arbitrary modality availability spanning multi-sequence MRI and amyloid-PET. A single model accepts whatever imaging is available, from a lone T1 scan to a full multimodal workup. Pretraining learns structural-molecular correspondences between MRI and PET via cross-modal distillation (RCMD) and prioritizes disease-vulnerable anatomy via atlas-guided curriculum masking (PACM), all within a shared 3D masked autoencoder (Multi-MAE3D). Across four downstream tasks and five clinically motivated modality settings, BrainAnytime largely outperforms modality-specific models, missing-modality baselines, and large-scale brain MRI pretrained foundation models on most modality settings. Notably, it surpasses the strongest missing-modality baselines with relative improvements of 6.2% and 7.0% in average accuracy on CN vs. AD and CN vs. MCI classification, respectively. Code is available at https://github.com/SDH-Lab/BrainAnytime.

Infrared and Visible Image Fusion (IVIF) has shown promise in visual tasks under challenging environments, but fusion under unregistered conditions faces inherent misalignments. Current studies to solve them either predict the deformation parameters coarse-to-fine (i.e., coarse registration and fine registration) or estimate the deformation fields in multi-scales for registration. Though straightforward, they overlook the cumulative errors in registration, which contaminate the fusion stage and severely deteriorate the resulting images. We introduce the Spatial-Frequency Registration and Fusion (SFRF) framework, which incorporates uncertainty estimation and infrared thermal radiation distribution consistency into a unified pipeline to handle the error accumulation for robust registration and fusion across both spatial and frequency domains. Specifically, SFRF constructs a Multi-scale Iterative Registration (MIR) framework that iteratively refines the deformation field across scales, leveraging uncertainty estimation at each stage to mitigate error accumulation and enhance alignment accuracy dynamically. To ensure the accurate alignment of infrared thermal distributions during registration, thermal radiation distribution consistency is employed as a frequency-domain supervisory signal, promoting global consistency in the frequency domain. Based on the spatial-frequency alignment, SFRF further adopts a Dual-branch Spatial-Frequency Fusion (DSFF) module, which incorporates spatial geometric features and frequency distribution information to reconstruct visually appealing images. SFRF achieves impressive performance across diverse datasets.

Evaluating whether large vision-language models (VLMs) align with human perception for high-level semantic scene comprehension remains a challenge. Traditional white-box interpretability methods are inapplicable to closed-source architectures and passive metrics fail to isolate causal features. We introduce Counterfactual Semantic Saliency (CSS). This black-box, model-agnostic framework quantifies the importance of objects by measuring the semantic shift induced by their causal ablation from a scene. To evaluate AI-human semantic alignment, we tested prominent VLMs against a human psychophysics baseline comprising 16,289 valid responses across 307 complex natural scenes and 1,306 high-fidelity counterfactual variants. Our analysis reveals a pervasive scene comprehension gap: models exhibit an overreliance (relative to humans) on large objects (size bias), objects at the center of the image (center bias), and high saliency objects. In contrast, models rely less on people in the scenes than our human participants to describe the images. A model's size bias is a primary driver explaining variations in model-human semantic divergence. Code and data will be available at https://github.com/starsky77/Counterfactual-Semantic-Saliency.

With recent advances in embodied agents and AR devices, egocentric observations are readily available as input for real-world interactive online applications. However, egocentric viewpoints can only sporadically observe hands, in addition to the estimated head trajectory. We propose EgoForce, an online framework for reconstructing long-term full-body motion from noisy egocentric input. While existing generative frameworks can robustly handle noisy and sparse measurements, they assume a fixed-length observation window is available and are thus not suitable for real-time applications. Faster inference often relies on autoregressive prediction, sacrificing robustness. In contrast, we adopt a diffusion-based method with a temporally asymmetric noise schedule inspired by Diffusion Forcing. Specifically, our approach models temporally evolving uncertainty and incrementally denoises states as new streaming observations arrive. Combined with a noise-robust imputation strategy, EgoForce progressively generates stable and coherent full-body motion under strict causal constraints. Experiments demonstrate that our online framework outperforms existing online and offline methods, enabling long-horizon, full-body motion reconstruction in challenging egocentric scenarios.

Geometric estimation including depth estimation and scene reconstruction is a crucial technique for colonoscopy which can provide surgeons with 3D spatial perception and navigation. However, geometric ground truth in colonoscopy is difficult to obtain due to narrow and enclosed space of the colon, while there is a large feature gap between simulated data and realistic data caused by artifacts and illumination. In this paper, we present CoGE, a novel framework for online monocular geometric estimation during colonoscopy. Firstly, we propose an illumination-aware supervision module based on the Retinex theory to address illumination diversity in different colonoscopy scenes. Moreover, a structure-aware perception module is proposed based on wavelet decomposition to extract common structural and local features of the colon. Both quantitative and qualitative results demonstrate that the proposed model solely trained on simulated data achieves state-of-the-art performance in geometric estimation for both simulated and realistic scenes.

Recent deep research systems have improved the ability of large language models to produce long, grounded reports through iterative retrieval and reasoning. However, most text-centered systems rely mainly on textual evidence, while multimodal systems often retrieve images only weakly or generate charts themselves, leaving source figures underused as evidence. We present ViDR, a multimodal deep research framework that grounds long-form reports in source figures. ViDR treats source figures as retrievable, interpretable, routable, and verifiable evidence objects, while still generating analytical charts when needed. It builds an evidence-indexed outline linking claims to textual and visual evidence, refines noisy web images into source-figure evidence atoms through context-aware filtering, outline-aware reranking, and VLM-based visual analysis, and generates each section with section-specific evidence. ViDR further validates visual references to reduce hallucinated or misplaced figures. We also introduce MMR Bench+, a benchmark for evaluating visual evidence use in deep research reports, covering source-figure retrieval, placement, interpretation, verifiability, and analytical chart generation. Experiments show that ViDR improves overall report quality, source-figure integration, and verifiability over strong commercial and open-source baselines. These results suggest that source visual evidence is important for multimodal deep research, as it strengthens evidential grounding, visual support, and report verifiability.

Text image super-resolution (Text-SR) requires more than visually plausible detail synthesis: slight errors in stroke topology may alter character identity and break readability. Existing methods improve text fidelity with stronger recognition-based or generative priors, yet they still face two unresolved challenges under severe degradation: the text condition extracted from low-quality inputs can itself be unreliable, and a plausible global prior does not fully determine fine-grained stroke boundaries. We present PRISM, a single-step diffusion-based Text-SR framework that addresses these two challenges through Flow-Matching Prior Rectification (FMPR) and a Structure-guided Uncertainty-aware Residual Encoder (SURE). FMPR constructs a privileged training-time prior from paired low-quality/high-quality latents and learns a flow matching that transports degraded embeddings toward this restoration-oriented prior space, yielding more accurate and reliable global text guidance. SURE further predicts uncertainty-aware structural residuals to selectively absorb reliable local boundary evidence while suppressing ambiguous stroke cues. Together, these components enable explicit global prior rectification and local structure refinement within a single diffusion restoration pass. Experiments on both synthetic and real-world benchmarks show that PRISM achieves state-of-the-art performance with millisecond-level inference. Our dataset and code will be available at https://github.com/faithxuz/PRISM.

Object-centric scene understanding is a fundamental challenge in computer vision. Existing approaches often rely on multi-stage pipelines that first apply pre-trained segmentors to extract individual objects, followed by per-object 3D reconstruction. Such methods are computationally expensive, fragile to segmentation errors, and scale poorly with scene complexity. We introduce OCH3R, a unified framework for Object-Centric Holistic 3D Reconstruction from a single RGB image. OCH3R performs one forward pass to simultaneously predict all object instances with their 6D poses and detailed 3D reconstructions. The key idea is a transformer architecture that predicts per-pixel attributes, including CLIP-based category embeddings, metric depth, normalized object coordinates (NOCS), and a fixed number of 3D Gaussians representing each object. To supervise these Gaussian reconstructions, we transform them into canonical space using the predicted 6D poses and align them with pre-rendered canonical ground truth, avoiding costly per-image Gaussian label generation. On standard indoor benchmarks, OCH3R achieves state-of-the-art performance across monocular depth estimation, open-vocabulary semantic segmentation, and RGB-only category-level 6D pose estimation, while producing high-fidelity, editable per-object reconstructions. Crucially, inference is fully feed-forward and scales independently of the number of objects, offering orders-of-magnitude speedups over conventional multi-stage pipelines in cluttered scenes.

The geometry of the retinal vessel is a key biomarker of vascular diseases, yet clinical evidence remains primarily observational. Existing generative counterfactuals intervene only at the image-level disease label, failing to isolate explicit anatomical structure. To address this limitation, we propose the Bézier Tree Encoding Counterfactual Framework (BTECF). By abstracting vascular networks into interconnected cubic-Bézier segments, BTECF establishes a disease-agnostic representation in which structural topology is explicitly preserved and atomically perturbable. Coupling this encoding with a diffusion-based generator enables parameter-level do-interventions on explicit geometric axes (e.g., tortuosity, caliber) while preserving background fundus textures. We validate BTECF on diabetic retinopathy, together with independent cohorts for ischemic stroke and Alzheimer's disease. Isolated counterfactual interventions produce dose-responsive shifts in classifier predictions; a matched pixel-drop control attenuates this response by an order of magnitude or more, ruling out out-of-distribution generation artifacts. By enforcing causal isolation between vessel topology and pixel-level confounders, BTECF provides a unified generative paradigm for hypothesis verification across systemic diseases. To support reproducibility, the code will be publicly released upon acceptance.

We study image inpainting with generative diffusion models. Existing methods typically either train dedicated task-specific models, or adapt a pretrained diffusion model separately for each masked image at deployment. We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We formulate it as a deterministic guidance problem with a supervised terminal objective. To make this problem learnable in high dimensions, we derive an auxiliary Gaussian formulation and prove that solving this randomized problem recovers the optimal deterministic guidance field. This bridge yields a principled continuous-time actor--critic algorithm for learning the guidance module in a fully data-driven manner. Empirically, on AFHQv2 and FFHQ under the pixel EDM pipeline and on ImageNet under the latent EDM2 pipeline, AID consistently improves the quality--speed trade-off over strong fixed-backbone and amortized inpainting baselines across multiple mask types, while adding less than one percent trainable overhead.

The rapid advancement of generative AI has enabled the creation of highly realistic and diverse synthetic images, posing critical challenges for image provenance and misinformation detection. This underscores the urgent need for effective image attribution. However, existing attribution datasets are constrained by limited scale, outdated generation methods, and insufficient semantic diversity - hindering the development of robust and generalizable attribution models. To address these limitations, we introduce ImageAttributionBench, a comprehensive dataset comprising images synthesized by a wide array of advanced generative models with state-of-the-art (SOTA) architectures. Covering multiple real-world semantic domains, the dataset offers rich diversity and scale to support and accelerate progress in image attribution research. To simulate real-world attribution scenarios, we evaluate several SOTA attribution methods on ImageAttributionBench under two challenging settings: (1) training on a standard balanced split and testing on degraded images, and (2) training and testing on semantically disjoint splits. In both cases, current methods exhibit consistently poor performance, revealing significant limitations in their robustness and generalization to unseen semantic content. Our work provides a rigorous benchmark to facilitate the development and evaluation of future image attribution methods.

Recent developments in generative models and large-scale datasets have substantially advanced 3D world generation, facilitating a broad range of domains including spatial intelligence, embodied intelligence, and autonomous driving. While achieving remarkable progress, existing approaches to 3D world generation typically prioritize appearance prediction with limited modeling of the underlying geometry, leading to issues such as unreliable scene structure estimation and degraded cross-view consistency. To address these limitations, motivated by the coarse-to-fine nature of human visual perception, we propose GTA, a novel image-to-3D world generation method following a Geometry-Then-Appearance paradigm. Specifically, given a single input image, to improve the structural fidelity of synthesized 3D scenes, GTA adopts a two-stage framework with two dedicated video diffusion models, which first generate coarse geometric structure from novel viewpoints and then synthesize fine-grained appearance conditioned on the predicted geometry. To further enhance cross-view appearance consistency, we introduce a random latent shuffle strategy during the training process, along with a test-time scaling scheme that improves perceptual quality without compromising quantitative performance. Extensive experiments have demonstrated that our proposed method consistently outperforms existing approaches in terms of fidelity, visual quality, and geometric accuracy. Moreover, GTA is shown to be effective as a general enhancement module that further improves the generation quality of existing image-to-3D world pipelines, as well as supporting multiple downstream applications and exhibiting favorable data efficiency during model training, highlighting its versatility and broad applicability. Project page: https://hanxinzhu-lab.github.io/GTA/.

Long video understanding is heavily bottlenecked by a rigid one-shot paradigm: existing methods either densely encode videos at prohibitive memory and latency costs, or aggressively compress them into sparse frame sets that irreversibly discard fine-grained evidence needed for downstream reasoning. Consequently, current models struggle to simultaneously balance temporal coverage, visual details, and computational efficiency. We propose AdaFocus, an efficient framework that rethinks long-video understanding as progressive evidence acquisition rather than one-pass encoding. AdaFocus relies on two tightly coupled components. First, a Query-Aware Adaptive Relevance-Diversity sampler (AdaRD) produces a compact yet informative video preview, adaptively switching to global clustering when the query lacks reliable local grounding. Second, instead of caching exhaustive frame sequences in memory, AdaFocus introduces an uncertainty-triggered refinement mechanism. It performs targeted look-back only when the model is not confident, retrieving high-resolution evidence directly from disk via a zero-cache I/O design. This turns discarded visual details from an irreversible loss into on-demand recoverable evidence without paying the cost of exhaustive preloading. Experiments on seven standard long-video benchmarks show that AdaFocus delivers a substantially better efficiency-accuracy trade-off than strong baselines. Compared with conventional dense encoding, AdaFocus achieves improved task performance (e.g., +2.59 accuracy on VideoMME, +8.39 mIoU on Charades-STA over single-pass inference) while reducing visual token consumption by ~33x and eliminating the need for in-memory frame pre-caching through its zero-cache disk retrieval design. These findings suggest that progressive preview combined with zero-cache evidence refinement is a highly effective paradigm for scalable multimedia reasoning.

Language-guided segmentation transcends the scope limitations of traditional semantic segmentation, enabling models to segment arbitrary target regions based on natural language instructions. Existing approaches typically adopt a two-stage framework: employing Multimodal Large Language Models (MLLMs) to interpret instructions and generate visual prompts, followed by foundational segmentation models (e.g., SAM) to produce masks. However, due to the limited spatial grounding capabilities of off-the-shelf MLLMs, these methods often rely on extensive training on large-scale datasets to achieve satisfactory accuracy. While recent advances have introduced reasoning mechanisms to improve performance, they predominantly operate within the textual domain, performing chain-of-thought reasoning solely based on abstract text representations without direct visual feedback. In this paper, we propose Seg-Agent, a completely training-free framework that pioneers Explicit Multimodal Chain-of-Reasoning. Unlike prior text-only reasoning, our approach constructs an interactive visual reasoning loop comprising three stages: generation, selection, and refinement. Specifically, we leverage Set-of-Mark (SoM) visual prompting to render candidate regions directly onto the image, allowing the MLLM to ``see'' and iteratively reason about spatial relationships in the visual domain rather than just the textual one. This explicit multimodal interaction enables Seg-Agent to achieve performance comparable to state-of-the-art training-based methods without any parameter updates. Furthermore, to comprehensively evaluate generalization across diverse scenarios, we introduce Various-LangSeg, a novel benchmark covering explicit semantic, generic object, and reasoning-guided segmentation tasks. Extensive experiments demonstrate the effectiveness and robustness of our method.

Grad-ECLIP is published at ICML 2024 and represents a new Transformer interpretation technical route (intermediate features-based). First, this paper demonstrates that the intermediate features-based technical route is not a novel one. Based on the existing attention-based route, we have developed Attention-ECLIP, which is completely equivalent to Grad-ECLIP but with simpler computation. Both through formal derivation and experimental validation, we prove that the intermediate feature-based route represented by Grad-ECLIP is actually an equivalent variant of the attention-based route. Next, this paper demonstrates that the Grad-ECLIP method is flawed. The model interpretation results obtained by Grad-ECLIP are not those of the original model, and the interpretation results are misaligned with the model's performance. We analyze the causes of Grad-ECLIP's flaws and propose, or rather, explicitly emphasize two fundamental principles that model interpretation should adhere to in order to avoid similar errors.

Recent diffusion- and flow-based VTON methods achieve strong results with pretrained generative models, but their reliance on multi-step sampling incurs high inference cost, while existing acceleration methods largely overlook the intrinsic structure of the try-on task. In this paper, we highlight a key observation: VTON outputs are highly constrained by the conditional inputs, suggesting that the conditional sampling trajectory can be much straighter than that in general image generation, making one-step generation a natural solution. However, limited task-specific data makes training from scratch impractical, forcing existing methods to fine-tune pretrained models whose objectives do not encourage such straight conditional trajectories. Thus, the deviation from an ideal straight path mainly comes from the mismatch between pretrained base models and the conditional nature of try-on generation, rather than from the task itself. Motivated by this insight, we encourage straighter VTON sampling trajectories through three targeted modifications: pure conditional transport, a garment preservation loss, and a self consistency loss. We further introduce a one-step distillation stage. Extensive experiments show that our method achieves state-of-the-art performance with one-step sampling, establishing a new standard for efficient and high-quality VTON.

Camera-conditioned video generation requires positional encoding that remains reliable under changes in camera motion, lens configuration, and scene structure. However, existing attention-level camera encodings either provide ray-only camera signals or rely on pinhole camera geometry, limiting their applicability to general camera control under the Unified Camera Model, including wide-angle and fisheye lenses. To address this limitation, we propose Curved Ray Expectation Positional Encoding (CRePE). CRePE represents each image token as a depth-aware positional distribution along its source ray, providing a Unified Camera Model-compatible positional encoding that captures the projected-path geometry induced by wide-angle and fisheye cameras. CRePE is implemented through a Geometric Attention Adapter added to frozen video DiTs, injecting token-wise scene-distance information into selected attention layers and stabilizing it with pseudo supervision from a monocular geometry foundation model. This design leads to more stable camera control and improves several geometry-aware and perceptual-quality metrics, while remaining competitive on video-quality metrics. Controlled positional-encoding ablations show a better overall average rank than a RayRoPE-style endpoint PE baseline, demonstrating the effectiveness of UCM-aware projected-path integration across diverse camera models. Furthermore, by extending the same positional-encoding pathway to external geometry control through Radial MixForcing, CRePE supports external radial-map control for scene-geometry-conditioned generation and source-video motion transfer beyond camera control.

Anti-facial recognition (AFR) image filters alter images in ways that are subtle to people but blinding to computer vision. Yet, despite widespread interest in these technologies to subvert surveillance, users rarely use them in practice -- because the ``subtle'' alterations are visible enough to conflict with users' self-presentation goals. To address this challenge, we propose AuraMask: a novel approach to creating AFR filters that are both adversarially effective and aesthetically acceptable. Using AuraMask, we produce 40 ``aesthetic'' filters that emulate popular ``one-click'' Instagram image filters. We show that AuraMask filters meet or exceed the adversarial effectiveness of prior methods against open-source facial recognition models. Moreover, in a controlled online user study ($N=630$) we confirm these filters achieve significantly higher user acceptance than prior methods. Lastly, we provide our AFR pipeline to the community for accelerated research in adversarially effective and aesthetically acceptable protections.

Standalone virtual reality (VR) headsets process highly sensitive personal, professional, and health-related data, yet their susceptibility to non-contact physical side channels remains largely unexplored. Existing side-channel attacks typically require malicious software execution or physical access to peripherals, making them conspicuous and potentially patchable. This paper introduces ThermalTap, the first passive, non-contact side-channel attack that fingerprints VR applications solely from the long-wave infrared (LWIR) radiation emitted by the headset chassis. By treating a headset's thermal signature as a high-fidelity proxy for internal computational workloads, ThermalTap enables remote application inference at meter-scale distances without any device interaction. To achieve robust performance in real-world settings, the system combines a commodity thermal camera with a multi-modal sensor suite (capturing ambient temperature, humidity, and airflow) to normalize environmental noise. We evaluate ThermalTap using six applications across three commercial standalone headsets. In indoor settings, ThermalTap identifies applications with over 90% accuracy using only 10 seconds of thermal camera data. Under outdoor conditions, with longer session-level observations, several applications remain identifiable despite environmental variability, with the strongest outdoor application reaching 81% accuracy. Our findings establish thermal radiation as a fundamental and unavoidable privacy risk for immersive systems, exposing a critical security gap that bypasses current software-level protections and physical access controls.

3D Gaussian Splatting (3DGS) is becoming a practical representation for novel view synthesis, but its growing adoption, together with rapid advances in instruction-driven 3DGS editing, also exposes a dual copyright risk: once a 3DGS-based asset is released, it can be used without permission and manipulated through 3D editing. Existing protection methods address only one side of this problem. Watermarking can trace ownership after unauthorized use, but it cannot prevent malicious editing. Adversarial edit-deterrence methods can disrupt editing, but they do not provide evidence of ownership. To the best of our knowledge, we present the first unified protection framework for 3DGS that jointly optimizes ownership tracing and unauthorized editing deterrence. Our framework combines a scene-wide watermarking objective over all Gaussians with an adversarial objective for edit deterrence. The adversarial branch combines latent-anchor separation, denoising-trajectory diversion, and cross-attention diversion to divert the editing trajectory, while an update-saliency-motivated Gaussian selection strategy assigns stronger adversarial updates to mask-selected Gaussians, improving the balance among watermark recovery, edit deterrence, and rendering fidelity. Experiments on scenes from Mip-NeRF 360 and Instruct-NeRF2NeRF demonstrate that the proposed framework achieves a favorable balance among bit accuracy, edit deterrence, and rendering quality. These results suggest that practical copyright protection of 3DGS-based assets can be more effectively addressed by integrating ownership tracing and unauthorized editing deterrence into a single optimization framework.

Deep learning models for medical imaging often exhibit overconfidence, creating safety risks in ambiguous diagnostic scenarios. While Conformal Prediction (CP) provides distribution-free statistical guarantees, standard methods such as Regularized Adaptive Prediction Sets (RAPS) optimize for average efficiency and can mask severe failures on difficult inputs. We propose an Adaptive Lambda Criterion for RAPS that minimizes the worst-case coverage violation across prediction set size strata. On OrganAMNIST (58,850 abdominal CT images, 11 classes), standard size-optimized RAPS converges to near-deterministic behavior with stratified undercoverage on uncertain samples, while our method achieves 95.72 percent global coverage with average set size 1.09 and at least 90 percent coverage across all strata. Cross-domain validation on PathMNIST (107,180 pathology images, 9 classes) confirms generalizability. Quantitative Grad-CAM analysis (rho = -0.30, p < 1e-22) shows that multi-label predictions correspond to focused attention on anatomically ambiguous regions. These results demonstrate that the proposed method improves reliability while maintaining efficiency, making it suitable for safety-critical medical AI applications.

Multimodal Large Language Models (MLLMs) have significantly advanced document understanding, yet current Doc-VQA evaluations score only the final answer and leave the supporting evidence unchecked. This answer-only approach masks a critical failure mode: a model can land on the correct answer while grounding it in the wrong passage -- a critical risk in high-stakes domains like law, finance, and medicine, where every conclusion must be traceable to a specific source region. To address this, we introduce CiteVQA, a benchmark that requires models to return element-level bounding-box citations alongside each answer, evaluating both jointly. CiteVQA comprises 1,897 questions across 711 PDFs spanning seven domains and two languages, averaging 40.6 pages per document. To ensure fidelity and scalability, the ground-truth citations are generated by an automated pipeline-which identifies crucial evidence via masking ablation-and are subsequently validated through expert review. At the core of our evaluation is Strict Attributed Accuracy (SAA), which credits a prediction only when the answer and the cited region are both correct. Auditing 20 MLLMs reveals a pervasive Attribution Hallucination: models frequently produce the right answer while citing the wrong region. The strongest system (Gemini-3.1-Pro-Preview) achieves an SAA of only 76.0, and the strongest open-source MLLM reaches just 22.5. Ultimately, towards trustworthy document intelligence, CiteVQA exposes a reliability gap that answer-only evaluations overlook, providing the instrumentation needed to close it. Our repository is available at https://github.com/opendatalab/CiteVQA.

Clinical trial studies indicate benefit of watch-and-wait (WW) surveillance for patients with rectal cancer showing a complete or near clinical response (CR) directly after treatment (restaging). However, there are no objectively accurate methods to early detect local tumor regrowth (LR) in patients undergoing WW from follow-up exams. Hence, we developed Temporal Rectal Endoscopy Cross-attention (TREX), a longitudinal deep learning approach that combines pairs of images acquired at restaging and follow-up to distinguish CR from LR. TREX uses pretrained Swin Transformers in a siamese setting to extract features from longitudinal images and dual cross-attention to combine the features without spatial co-registration between image pairs. TREX and Swin-based baselines were trained under two settings: (a) detecting LR or CR at the last available follow-up and (b) early detection of LR at 3--6, 6--12, and 12--24 months before clinical confirmation. TREX achieved the highest accuracy in detecting LR with a high sensitivity of 97% $\pm$ 6% and a balanced accuracy of 90% $\pm$ 3%, and outperformed all baselines in early detection at both 3--6 (74% $\pm$ 1%) and 6--12 months (62% $\pm$ 4%) prior to clinical detection. Clinical validation via a surgeon survey showed that TREX matched attending-level overall accuracy (TREX: 86.21% vs.\ Clinicians: 87.84% $\pm$ 1.28%). Finally, we explored TREX's ability to predict treatment response by combining pre-treatment (pre-TNT) and restaging endoscopies, achieving a balanced accuracy of 73% $\pm$ 12%. These results show that longitudinal deep learning analysis of endoscopy may improve surveillance and enable earlier identification of rectal cancer regrowth.

Automated analysis of peripheral blood smears for Acute Lymphoblastic Leukemia (ALL) is hindered by low contrast and substantial variability in cytoplasmic appearance, which complicate conventional membrane-based segmentation. We found that many recent approaches rely on heavy neural architectures and extensive training, but still struggle to generalize across staining and acquisition variability. To address these limitations, we propose the Perinuclear Ring-based Image Segmentation Method (PRISM), which replaces explicit cytoplasmic delineation with adaptive concentric zones constructed around the nucleus. These perinuclear regions enable the extraction of robust cytoplasmic descriptors by integrating color information with texture statistics derived from grey-level co-occurrence patterns, without requiring accurate cell-boundary detection. A calibrated stacking ensemble of traditional classifiers leverages these descriptors to achieve a high performance, with an accuracy of 98.46% and a precision-recall AUC of 0.9937.

Assembling objects from parts requires understanding multimodal instructions, linking them to 3D components, and predicting physically plausible 6-DoF motions for each assembly step. Existing datasets focus on simplified scenarios, overlooking shape complexities and assembly trajectories in industrial assemblies. We introduce AssemblyBench, a synthetic dataset of 2,789 industrial objects with multimodal instruction manuals, corresponding 3D part models, and part assembly trajectories. We also propose a transformer-based model, AssemblyDyno, which uses the instructional manual and the 3D shape of each part to jointly predict assembly order and part assembly trajectories. AssemblyDyno outperforms prior works in both assembly pose estimation and trajectory feasibility, where the latter is evaluated by our physics-based simulations.

The proliferation of sophisticated image editing tools and generative artificial intelligence models has made verifying the authenticity of digital images increasingly challenging, with important implications for journalism, forensic analysis, and public trust. Although numerous forensic algorithms, ranging from handcrafted methods to deep learning-based detectors, have been developed for manipulation detection, individual methods often suffer from limited robustness, fragmented evidence, or weak generalization across manipulation types and image conditions. To address these limitations, we present \textbf{FRAME}, a method for \textbf{F}orensic \textbf{R}outing and \textbf{A}daptive \textbf{M}ulti-path \textbf{E}vidence fusion for image manipulation detection. FRAME organizes diverse forensic algorithms into a multi-path analysis space, adaptively selects informative forensic paths for each input image, and fuses complementary evidence to improve detection and localization performance. By moving beyond single-method analysis and fixed fusion strategies, FRAME provides a more robust and flexible approach to image forensic reasoning while preserving interpretable forensic cues from multiple evidence sources. Experimental results demonstrate the effectiveness of FRAME across diverse manipulation scenarios. Code is available at \href{https://github.com/kzhao5/FRAME}{https://github.com/kzhao5/FRAME}.

Recent advances in generative models have yielded impressive progress on motion in-betweening, allowing for more complex, varied, and realistic motion transitions. However, recent methods still exhibit noticeable limitations in preserving keyframe information and ensuring motion continuity. In this paper, we propose a novel pipeline and sampling optimization strategy for latent diffusion models (LDM) based on motion implicit neural representations (INR). By establishing a mapping between INR and sparse spatial or temporal information within latent diffusion, our model can sample the INR parameters from extremely sparse and ambiguous keyframe data and reconstruct plausible and smooth motions from the manifold. Our experiments demonstrate the superior performance of our model, which significantly improves motion generation quality in scenarios with few keyframes while ensuring both keyframe accuracy and diversity of in-between motions.

Estimating camera pose in dynamic environments is a critical challenge, as most visual SLAM and SfM methods assume static scenes. While recent dynamic-aware methods exist, they are often not unified: semantic-based approaches are brittle, per-sequence optimization methods fail on short sequences, and other learned models may degrade on static-only scenes. We present WildPose, a unified monocular pose estimation framework that is robust in dynamic environments while maintaining state-of-the-art performance on static and low-ego-motion datasets. Our key insight is to connect two powerful paradigms in modern 3D vision: the rich perceptual frontend of feedforward models and the end-to-end optimization of differentiable bundle adjustment (BA). We achieve this with a 3D-aware update operator built on a frozen, pre-trained MASt3R feature backbone, together with a high-capacity motion mask detector that uses multi-level 3D-aware features from the same backbone. Extensive experiments show WildPose consistently outperforms prior methods across dynamic (Wild-SLAM, Bonn), static (TUM, 7-Scenes), and low-ego-motion (Sintel) benchmarks.

Wu et al. (2026) showed that most frontier large language models (LLMs) recommend a sponsored, roughly twice-as-expensive flight when their system prompt contains a soft sponsorship cue. We reproduce their evaluation on ten open-weight chat models plus the two of their twenty-three models that are still reachable today (gpt-3.5-turbo, gpt-4o). All reported rates in this paper are produced under the same judge the original paper used (gpt-4o); we additionally store every label under an open-weight (gpt-oss-120b) and a smaller proprietary (gpt-4o-mini) judge for an ablation. Three findings emerge. First, a prose description of an LLM evaluation pipeline is not, on its own, sufficient for accurate reproduction: we surfaced three silent implementation failures that each shifted a reported rate by tens of percentage points. Second, the central claims do generalise - the gpt-3.5-turbo logistic-regression intercept of alpha = 0.81 is within four points of the original alpha = 0.86, and 200 of 200 trials on gpt-3.5-turbo and gpt-4o promote a payday lender to a financially distressed user. Third, a thirty-token user prompt that asks the assistant for a neutral comparison table first cuts sponsored recommendation from 46.9% to 1.0% averaged across our ten open-source models, and from 53.0% to 0% averaged across the two OpenAI models. AI literacy and price-comparison portals are likely market-level mitigations; the harmful-product cell is bounded by neither. Raw data, labels and analysis scripts are at https://github.com/akmaier/Paper-LLM-Ads .

INTRODUCTION | Fully supervised 3D segmentation of high-resolution ex vivo MRI is limited by the prohibitive cost of volumetric annotation, forcing reliance on sparse 2D slices. Weakly supervised Sparse-to-Dense frameworks bridge this gap, but guidelines remain ambiguous regarding human-centric visual enhancements and transferring optimization strategies across dimensions. We analyze divergent regularization needs for multi-class segmentation of high-resolution ex vivo spinal cord MRI. METHODS | We used 9.4T MRI of multiple sclerosis spinal cords (>104,000 slices) with sparse annotations (428 slices). A 2D Teacher trained on sparse slices generated dense pseudo-labels to train a 3D Student. We systematically evaluated the impact of human-centric preprocessing, spatial augmentation, and soft-label regularization on both architectures. RESULTS | We identified a critical divergence in training dynamics. The 2D Teacher required strong spatial augmentation and soft-labeling to overcome data scarcity, improving White Matter Lesion Dice scores by >11 points. However, propagating these techniques to the 3D Student degraded its performance. Furthermore, human-centric preprocessing (e.g., CLAHE) disrupted global statistical cues, dropping Gray Matter Lesion Dice scores by ~25 points. DISCUSSION | Our study highlights a perception divergence (human-centric contrast enhancement harms machine models) and a regularization conflict across dimensions. 3D architectures trained on dense pseudo-labels exhibit fundamentally different optimization landscapes than 2D counterparts and require distinct, conservative regularization. Code and models: https://github.com/ivadomed/model_seg_sc-gm-lesion_human_ms_exvivo_t2star.

Existing physical adversarial attacks on vision-based autonomous driving induce time-evolving perception errors, including biased object tracking or trajectory prediction, through (i) sophisticated physical patch inducing detection box drift when entering the view distance, or (ii) dynamically changing patches that cause different perception errors at different time. In both cases, viewing-angle variation is treated as a challenge, requiring adversarial patches to remain effective across frames under varying views, leading to complex multi-view optimization. In contrast, we show that viewing-angle variation itself can be turned into an attack tool. We design a new attack paradigm where a static, passive adversarial camouflage is mounted on a vehicle whose view-dependent appearance naturally evolves with relative motion, inducing consistent feature drift across frames. This causes the system to infer a physically plausible but incorrect trajectory, such as a false cut-in, which propagates to downstream decision-making and triggers unnecessary braking. Unlike prior approaches that require multi-view robustness or active intervention, our attack emerges from normal driving dynamics and is easy to deploy: a parked vehicle with a natural camouflage can induce hard braking in passing autonomous vehicles. We demonstrate the novel attack on nuScenes dataset, showing the effectiveness with an end-to-end success rate of up to 87.5%, measured by hard-braking events, and robustness across different scene backgrounds, victim vehicle speeds, and perception models.

Recent video anomaly detection research has expanded rapidly with an emphasis on general models of normality intended to work across many different scenes. While this focus has led to improvements in scalability and multi-scene generalization, it has also shifted the field away from modeling the scene-specific and context-dependent nature of normal behavior. Contemporary approaches frequently rely on video-level weak supervision and opaque pretrained representations from multi-modal large language models (MLLMs), which encourage models to respond to familiar semantic anomaly categories rather than to deviations from the normal patterns of a particular environment. This trend suppresses spatial localization, introduces semantic bias, and reduces anomaly detection to a form of action recognition. In this paper, we examine whether these prevailing formulations align with the core requirements of real-world VAD, which is typically performed within a single scene where normality is determined by local geometry, semantics, and activity patterns. Through targeted visual analyses and empirical evaluations, we demonstrate the practical consequences of these limitations and show that meaningful progress in VAD requires renewed focus on single-scene, spatially-aware, and explainable formulations that capture the nuanced structure of normality within individual environments.

Instruction-based image editing exhibits heterogeneous difficulty not only across cases but also across regions of an image, motivating refinement approaches that allocate correction to where the model struggles. Existing refinement signals arrive late, after a fully generated image or a completed denoising step. We ask whether such a signal can act within an ongoing forward pass. To investigate this, we probe a frozen image-editing model and find that although generation capability emerges only in the last few layers, the error pattern is already set in early layers (rank correlation \r{ho} = 0.83 with the final-layer error map). Based on this, we introduce Inline Critic, a learnable token that critiques a frozen model's predictions at its intermediate layers and steers its hidden states to refine generation during the forward pass. A three-stage recipe is proposed to stabilize the training from learning how to critique to steering generation. As a result, we achieve state of the art on GEdit-Bench (7.89), a +9.4 gain on RISEBench over the same backbone, and the strongest open-source result on KRIS-Bench (81.92, surpassing GPT-4o). We further provide analyses showing that the critic genuinely shapes the model's attention and prediction updates at subsequent layers.

We introduce MMCL-Bench, a benchmark for multimodal context learning: learning task-local rules, procedures, and empirical patterns from visual or mixed-modality teaching context and applying them to new visual instances. Unlike text-only context learning or standard multimodal question answering, this setting requires models to recover and localize relevant evidence from images, screenshots, manuals, videos, and frame sequences before they can reason over the learned context. MMCL-Bench contains 102 tasks spanning three categories: rule system application, procedural task execution, and empirical discovery and induction. We evaluate frontier multimodal models with strict rubric-based scoring and find that current systems remain far from robust multimodal context learning, with even the strongest model solving fewer than one-third of tasks under strict evaluation. Diagnostic ablations and error analysis show that failures arise throughout the context-to-answer pipeline, including context anchoring, visual evidence extraction, context reasoning, and response construction. MMCL-Bench thus highlights multimodal context learning as an important unsolved capability bottleneck for current multimodal models.

Multimodal large language models (MLLMs) are now routinely deployed for visual understanding, generation, and curation. A substantial fraction of these applications require an explicit aesthetic judgment. Most existing solutions reduce this judgment to predicting a scalar score for a single image. We first ask whether such scores faithfully capture comparative preference: in a controlled study with eight expert annotators, score-derived rankings align poorly with the same annotators' direct comparisons, while direct ranking yields substantially higher inter-annotator agreement on best- and worst-image labels. Motivated by this finding, we introduce the Visual Aesthetic Benchmark (VAB), which casts aesthetic evaluation as comparative selection over candidate sets with matched subject matter. VAB contains 400 tasks and 1,195 images across fine art, photography, and illustration, with labels derived from the consensus of 10 independent expert judges per task. Evaluating 20 frontier MLLMs and six dedicated visual-quality reward models, we find that the strongest system identifies both the best and the worst image correctly across three random permutations of the candidate order in only 26.5% of tasks, far below the 68.9% achieved by human experts. Fine-tuning a 35B-parameter model on 2,000 expert examples brings its accuracy close to that of a 397B-parameter open-weight model, suggesting that the comparative signal in VAB is transferable. Together, these results expose a clear and measurable gap between current multimodal models and expert aesthetic judgment, and VAB provides the first set-based, expert-grounded testbed on which that gap can be tracked and closed.

Foundation diffusion models can generate photorealistic natural images, but adapting them to medical imaging remains challenging. In medical adaptation, limited labeled data can exacerbate hallucination-like and clinically implausible synthesis, while existing metrics such as FID or Inception Score do not quantify per-image alignment with pathology-relevant criteria. We introduce the Clinical Alignment Score (CAS), a foundation-model-based proxy for clinical alignment that evaluates generated images along four complementary dimensions beyond visual fidelity. Building on CAS, we propose Clinical Reward-Aligned Finetuning (CRAFT), a reward-based adaptation framework that transfers medical knowledge from multimodal large language models and vision-language models through label-conditioned prompt enrichment, clinical checklists, and differentiable reward optimization. Across four diverse modalities, CRAFT improves CAS and downstream classification performance over strong adaptation baselines. Beyond average CAS gains, CRAFT reduces the empirical low-alignment tail below a real-image reference threshold by 5.5-34.7% points relative to the strongest baseline, corresponding to a 20.4% average relative reduction across datasets. These results indicate fewer hallucination-like generations under CAS, and are corroborated by out-of-family evaluator evaluation, structured checklist auditing, memorization analysis, and a blinded physician preference study on CheXpert.

Pursuing training-free open-vocabulary semantic segmentation in an efficient and generalizable manner remains challenging due to the deep-seated spatial bias in CLIP. To overcome the limitations of existing solutions, this work moves beyond the CLIP-based paradigm and harnesses the recent spatially-aware dino$.$txt framework to facilitate more efficient and high-quality dense prediction. While dino$.$txt exhibits robust spatial awareness, we find that the semantic ambiguity of text queries gives rise to severe mismatch within its dense cross-modal interactions. To address this, we introduce Visual-guided Prompt evolution (VIP) to rectify the semantic expressiveness of text queries in dino$.$txt, unleashing its potential for fine-grained object perception. Towards this end, VIP integrates alias expansion with a visual-guided distillation mechanism to mine valuable semantic cues, which are robustly aggregated in a saliency-aware manner to yield a high-fidelity prediction. Extensive evaluations demonstrate that VIP: 1. surpasses the top-leading methods by 1.4%-8.4% average mIoU, 2. generalizes well to diverse challenging domains, and 3. requires marginal inference time and memory overhead.

In language reasoning, longer chains of thought consistently yield better performance, which naturally suggests that visual latent reasoning may likewise benefit from longer latent sequences. However, we discover a counterintuitive phenomenon: the performance of existing latent visual reasoning methods systematically degrades as the latent sequence grows longer. We reveal the root cause: Information Gain Collapse -- autoregressive generation makes each step highly dependent on prior outputs, so subsequent tokens can barely introduce new information. We further identify that heavily pooled ($\geq 128\times$) image embeddings used as supervision targets provide no more signal than meaningless placeholders. Motivated by these insights, we propose SCOLAR (Self-COnsistent LAtent Reasoning), which introduces a lightweight detransformer that leverages the LLM's full-sequence hidden states to generate auxiliary visual tokens in a single shot, with each token independently anchored to the original visual space. Combined with three-stage SFT and ALPO reinforcement learning, SCOLAR extends acceptable latent CoT length by over $30\times$, achieves state-of-the-art among open-source models on real-world reasoning benchmarks (+14.12% over backbone), and demonstrates strong out-of-distribution generalization.

Multimodal learning seeks to integrate information across diverse sensory sources, yet current approaches struggle to balance cross-modal generalizability with modality-specific structure. Continuous (implicit) methods preserve fine-grained priors but render generalization challenging, while discrete (explicit) approaches enforce shared prototypes at the expense of modality specificity. We introduce CoDAAR (Cross-modal Discrete Alignment And Reconstruction), a novel framework that resolves this long-standing trade-off by establishing semantic consensus across modality-specific codebooks through index-level alignment. This design uniquely allows CoDAAR to preserve modality-unique structures while achieving generalizable cross-modal representations within a unified discrete space. CoDAAR combines two complementary mechanisms: Discrete Temporal Alignment (DTA), which enables fine-grained temporal quantization, and Cascading Semantic Alignment (CSA), which promotes progressive cross-modal semantic agreement. Together, they establish a competition-free unified representation space. Trained with self-supervised reconstruction objectives on paired multimodal sequences, CoDAAR demonstrates robust cross-modal and cross-domain generalization. Across Cross-Modal Generalization benchmarks, including event classification, localization, video segmentation, and cross-dataset transfer, CoDAAR achieves state-of-the-art performance, establishing a new paradigm for discrete and generalizable multimodal representation learning.

Generative novel view synthesis faces a fundamental dilemma: geometric priors provide spatial alignment but become sparse and inaccurate under view changes, while appearance priors offer visual fidelity but lack geometric correspondence. Existing methods either propagate geometric errors throughout generation or suffer from signal conflicts when fusing both statically. We introduce MoCam, which employs structured denoising dynamics to orchestrate a coordinated progression from geometry to appearance within the diffusion process. MoCam first leverages geometric priors in early stages to anchor coarse structures and tolerate their incompleteness, then switches to appearance priors in later stages to actively correct geometric errors and refine details. This design naturally unifies static and dynamic view synthesis by temporally decoupling geometric alignment and appearance refinement within the diffusion process. Experiments demonstrate that MoCam significantly outperforms prior methods, particularly when point clouds contain severe holes or distortions, achieving robust geometry-appearance disentanglement.

Transfer learning is a machine learning technique that uses previously acquired knowledge from a source domain to enhance learning in a target domain by reusing learned weights. This technique is ubiquitous because of its great advantages in achieving high performance while saving training time, memory, and effort in network design. In this paper, we investigate how to select the best pre-trained model that meets the target domain requirements for image classification tasks. In our study, we refined the output layers and general network parameters to apply the knowledge of eleven image processing models, pre-trained on ImageNet, to five different target domain datasets. We measured the accuracy, accuracy density, training time, and model size to evaluate the pre-trained models both in training sessions in one episode and with ten episodes.

Audio-visual understanding requires effective alignment between heterogeneous modalities, yet cross-modal correspondence remains challenging when temporally aligned audio and visual signals lack clear semantic correspondence. We propose to use text as a semantic anchor for audio-visual representation learning. To this end, we introduce a parameter-efficient adaptation framework built on frozen audio and visual encoders, centered on Text-Bridged Audio-Visual Adapter (TB-AVA), which enables text-mediated interaction between audio and visual streams. At the core of TB-AVA, Gated Semantic Modulation (GSM) selectively modulates feature channels based on text-inferred semantic relevance. We evaluate the proposed approach on multiple benchmarks, including AVE, AVS, and AVVP, where the proposed framework achieves state-of-the-art performance, demonstrating text as an effective semantic anchor for parameter-efficient fine-tuning (PEFT) in audio-visual learning.

Clinical check-up reports are multimodal documents that combine page layouts, tables, numerical biomarkers, abnormality flags, imaging findings, and domain-specific terminology. Such heterogeneous evidence is difficult for laypersons to interpret and translate into concrete follow-up actions. Although large language models show promise in medical summarisation and triage support, their ability to generate safe, prioritised, and patient-oriented actions from multimodal check-up reports remains under-benchmarked. We present \textbf{Checkup2Action}, a multimodal clinical check-up report dataset and benchmark for structured \textit{Action Card} generation. Each card describes one clinically relevant issue and specifies its priority, recommended department, follow-up time window, patient-facing explanation, and questions for clinicians, while avoiding diagnostic or treatment-prescriptive claims. The dataset contains 2,000 de-identified real-world check-up reports covering demographic information, physical examinations, laboratory tests, cardiovascular assessments, and imaging-related evidence. We formulate checkup-to-action generation as a constrained structured generation task and introduce an evaluation protocol covering issue coverage and precision, priority consistency, department and time recommendation accuracy, action complexity, usefulness, readability, and safety compliance. Experiments with general-purpose and medical large language models reveal clear trade-offs between issue coverage, action correctness, conciseness, and safety alignment. Checkup2Action provides a new multimodal benchmark for evaluating patient-oriented reasoning over clinical check-up reports.

Adversarial attacks fool deep image classifiers by adding tiny, almost invisible noise patterns to a clean image. The standard $\ell^\infty$-bounded attacks (FGSM, PGD, and the $\ell^\infty$ variant of Carlini--Wagner) produce high-frequency, near-random sign patterns at the pixel level: nearly invisible in $\ell^2$, but carrying disproportionate gradient energy. We exploit this with a single-shot, training-free detector using the high-order Corbino--Castillo mimetic operators from the open-source MOLE library. No retraining, no surrogate classifier, no access to the network under attack: the verdict is a property of the input alone, computed in $O(HW)$ time. We validate the detector on the standard \texttt{peppers} test image at the canonical $\ell^\infty$ budget $\varepsilon = 16/255$ and observe a clean-vs-adversarial separation that grows monotonically from $3.55\times$ at order $k=2$ to $4.62\times$ at $k=8$.

All-in-one image restoration seeks to recover clean images from inputs affected by diverse and unknown degradations using a unified framework. Recent methods have shown strong performance by identifying degradation characteristics to guide the restoration process. However, many of them treat degradations as discrete categories, which limits their ability to model the continuous relational structure that arises in composite degradations. To address this issue, we propose a multimodal large language model (MLLM)-guided image restoration framework that exploits multimodal embeddings as guidance for low-level restoration. Specifically, MLLM-derived features are injected into an encoder-decoder architecture through an MLLM-guided fusion block (MGFB) to enhance degradation-aware representations. In addition, we incorporate a mixture-of-frequency-experts (MoFE) module that adaptively combines frequency experts using MLLM-guided contextual cues. To further improve expert routing, we design an MLLM-guided router with a relational alignment loss that encourages routing patterns consistent with the embedding-space relationships of degraded inputs. Extensive experiments on multiple benchmarks show that the proposed method achieves strong performance across diverse restoration settings and establishes a new state of the art on the challenging CDD11 dataset, outperforming previous methods by up to 1.35 dB.

Existing reward alignment methods for diffusion and flow models rely on multi-step stochastic trajectories, making them difficult to extend to deterministic generators. A natural alternative is noise-space optimization, but existing approaches require backpropagation through the generator and reward pipeline, limiting applicability to differentiable settings. To address this, here we present ZeNO (Zeroth-order Noise Optimization), a gradient-free framework that formulates noise optimization as a path-integral control problem, estimable from zeroth-order reward evaluations alone. When instantiated with an Ornstein--Uhlenbeck reference process, the update connects to Langevin dynamics implicitly targeting a reward-tilted distribution. ZeNO enables effective inference-time scaling and demonstrates strong performance across diverse generators and reward functions, including a protein structure generation task where backpropagation is infeasible.

Reinforcement learning (RL) has shown extraordinary potential in aligning diffusion models to downstream tasks, yet most of them still suffer from significant reward hacking, which degrades generative diversity and quality by inducing visual mode collapse and amplifying unreliable rewards. We identify the root cause as the mode-seeking nature of these methods, which maximize expected reward without effectively constraining probability distribution over acceptable trajectories, causing concentration on a few high-reward paths. In contrast, we propose Trajectory Matching Policy Optimization (TMPO), which replaces scalar reward maximization with trajectory-level reward distribution matching. Specifically, TMPO introduces a Softmax Trajectory Balance (Softmax-TB) objective to match the policy probabilities of K trajectories to a reward-induced Boltzmann distribution. We prove that this objective inherits the mode-covering property of forward KL divergence, preserving coverage over all acceptable trajectories while optimizing reward. To further reduce multi-trajectory training time on large-scale flow-matching models, TMPO incorporates Dynamic Stochastic Tree Sampling, where trajectories share denoising prefixes and branch at dynamically scheduled steps, reducing redundant computation while improving training effectiveness. Extensive results across diverse alignment tasks such as human preference, compositional generation and text rendering show that TMPO improves generative diversity over state-of-the-art methods by 9.1%, and achieves competitive performance in all downstream and efficiency metrics, attaining the optimal trade-off between reward and diversity.

Vision-language-action (VLA) models remain constrained by the scarcity of action-labeled robot data, whereas action-free videos provide abundant evidence of how the physical world changes. Latent action models offer a promising way to extract such priors from videos, but reconstruction-trained latent codes are not necessarily suitable for policy generation: they may predict future observations while lacking the structure needed to be reused or generated coherently with robot actions. We introduce ALAM (Algebraic Latent Action Model), an Algebraically Consistent Latent Action Model that turns temporal relations in action-free video into structural supervision. Given frame triplets, ALAM learns latent transitions that are grounded by reconstruction while being regularized by composition and reversal consistency, encouraging a locally additive transition space. For downstream VLA learning, we freeze the pretrained encoder and use its latent transition sequences as auxiliary generative targets, co-generated with robot actions under a joint flow-matching objective. This couples structured latent transitions with flow-based policy generation, allowing the policy to exploit ALAM's locally consistent transition geometry without requiring latent-to-action decoding. Representation probes show that ALAM reduces additivity and reversibility errors by 25-85 times over unstructured latent-action baselines and improves long-horizon cumulative reconstruction. When transferred to VLA policies, ALAM raises the average success rate from 47.9% to 85.0% on MetaWorld MT50 and from 94.1% to 98.1% on LIBERO, with consistent gains on real-world manipulation tasks. Ablations further confirm that the strongest improvements arise from the synergy between algebraically structured latent transitions and joint flow matching.

Vision-Language-Action (VLA) models have emerged as a promising paradigm for end-to-end autonomous driving. However, existing reasoning mechanisms still struggle to provide planning-oriented intermediate representations: textual Chain-of-Thought (CoT) fails to preserve continuous spatiotemporal structure, while latent world reasoning remains difficult to use as a direct condition for action generation. In this paper, we propose CoWorld-VLA, a multi-expert world reasoning framework for autonomous driving, where world representations serve as explicit conditions to guide action planning. CoWorld-VLA extracts complementary world information through multi-source supervision and encodes it into expert tokens within the VLA, thereby providing planner-accessible conditioning signals. Specifically, we construct four types of tokens: semantic interaction, geometric structure, dynamic evolution, and ego trajectory tokens, which respectively model interaction intent, spatial structure, future temporal dynamics, and behavioral goals. During action generation, CoWorld-VLA employs a diffusion-based hierarchical multi-expert fusion planner, which is coupled with scene context throughout the joint denoising process to generate continuous ego trajectories. Experiments show that CoWorld-VLA achieves competitive results in both future scene generation and planning on the NAVSIM v1 benchmark, demonstrating strong performance in collision avoidance and trajectory accuracy. Ablation studies further validate the complementarity of expert tokens and their effectiveness as planning conditions for action generation. Code will be available at https://github.com/AFARI-Research/CoWorld-VLA.

Scientific reasoning is a key aspect of human intelligence, requiring the integration of multimodal inputs, domain expertise, and multi-step inference across various subjects. Existing benchmarks for multimodal large language models (MLLMs) often fail to capture the complexity and traceability of reasoning processes necessary for rigorous evaluation. To fill this gap, we introduce SciVQR, a multimodal benchmark covering 54 subfields in mathematics, physics, chemistry, geography, astronomy, and biology. SciVQR includes domain-specific visuals, such as equations, charts, and diagrams, and challenges models to combine visual comprehension with reasoning. The tasks range from basic factual recall to complex, multi-step inferences, with 46% including expert-authored solutions. SciVQR not only evaluates final answers but also examines the reasoning process, providing insights into how models reach their conclusions. Our evaluation of leading MLLMs, including both proprietary and open-source models, reveals significant limitations in handling complex multimodal reasoning tasks, underscoring the need for improved multi-step reasoning and better integration of interdisciplinary knowledge in advancing MLLMs toward true scientific intelligence. The dataset and evaluation code are publicly available at https://github.com/CASIA-IVA-Lab/SciVQR.

Recent research work on fashion outfit generation focuses on promoting visual consistency of garments by leveraging key information from reference image and text prompt. However, the potential of outfit generation remains underexplored, requiring comprehensive e-commercial dataset and elaborative utilization of multi-modal condition. In this paper, we propose a brand-new e-commerce dataset, named Fashion130k, with various occasions, models, and garment types. For the consistent generation of garment, we design a framework with Unified Multi-modal Condition (UMC) to align and integrate the text and visual prompts into generation model. Specifically, we explore an embedding refiner to extract the unified embeddings of multi-modal prompts, within which a Fusion Transformer is proposed to align the multi-modal embeddings by adjusting the modality gap between text and image. Based on unified embeddings, the attention in generation model is redesigned to emphasis the correlations between prompts and noise image, inducing that the noise image can select the pivotal tokens of prompts for consistent outfit generation. Our dataset and proposed framework offer a general and nuanced exploration of multi-modal prompts for generation models. Extensive experiments on real-world applications and benchmark demonstrate the effectiveness of UMC in visual consistency, achieving promising result than that of SoTA methods.

The rapid evolution of generative paradigms has enabled the creation of highly realistic imagery, which escalating the risks of identity fraud and the dissemination of disinformation. Most existing approaches frame face forgery detection as a fully supervised binary classification problem. Consequently, these models typically exhibit significant performance decay when tasked with detecting forgeries from previously unseen generative paradigms. Furthermore, these methods focus exclusively on either DeepFakes or fully synthesized faces, thereby failing to provide a generalized framework for universal face forgery detection. In this paper, we address this challenge by introducing FADNet (Face Authenticity Detector Net), % a self-supervised framework that which reformulates face forgery detection as a one-class classification (OCC) task. By training exclusively on authentic facial data to capture their intrinsic representations, FADNet flags any image whose feature embedding deviates significantly from the learned distribution of real faces as a forgery. The framework incorporates Evidential Deep Learning (EDL) to quantify predictive uncertainty and utilizes a plug-and-play pseudo-forgery image generator (PFIG) to tighten decision boundaries around authentic data. Extensive experimental evaluations on the DF40 and ASFD benchmarks demonstrate that FADNet achieves superior performance and generalization capabilities. Specifically, FADNet substantially outperforms existing state-of-the-art (SOTA) methods, yielding a remarkable average accuracy of 96.63\% and an average precision of 98.83\%.

With the rapid development of deep generative models, forged facial images are massively exploited for illegal activities. Although existing synthetic face detection methods have achieved significant progress, they suffer from the inherent limitation of overconfidence due to their reliance on the Softmax activation function. Thus, these methods often lead to unreliable predictions when encountering unknown Out-of-Distribution (OOD) images, and cannot ascertain the model's uncertainty in its prediction. Meanwhile, most existing methods require massive high-quality annotated data, which greatly limits their practicability across diverse scenarios. To address these limitations, we propose EMSFD (Evidence-based decision Modeling for Synthetic Face Detection with uncertainty-driven active learning), an approach designed to enhance detection reliability and generalizability. Specifically, EMSFD models class evidence using the Dirichlet distribution and explicitly incorporates model uncertainty into the prediction process. Furthermore, during training, the estimated uncertainty is exploited to prioritize more informative samples from the unlabeled pool for annotation, thereby reducing labeling cost and improving model generalization. Extensive experimental evaluations demonstrate that our method enhances the interpretability of synthetic face detection. Meanwhile, our method yields a 15\% increase in accuracy compared to existing state-of-the-art (SOTA) baselines, which demonstrates the superior detection performance and generalizability of our approach. Our code is available at: https://github.com/hzx111621/EMSFD.

On-policy distillation (OPD), which supervises a student on its own sampled trajectories, has emerged as a data-efficient post-training method for improving reasoning while avoiding the reward dependence of reinforcement learning and the catastrophic forgetting often observed in standard supervised fine-tuning. However, standard OPD typically computes teacher supervision under noisy student-generated contexts and often relies on a single stochastic teacher rollout per prompt. As a result, the supervision signal can be high-variance: the sampled teacher trajectory can be incorrect, uninformative, or poorly matched to the student's current reasoning behavior. To address this limitation, we propose BRTS, a Best-of-N Rollout Teacher Selection framework for on-policy distillation. BRTS augments standard student-context OPD with a teacher-context supervision branch constructed from the curated teacher trajectory. Rather than distilling from the first sampled teacher rollout, BRTS samples a small pool of teacher trajectories and selects the auxiliary trajectory using a simple priority rule: correctness first, student alignment second. When multiple correct teacher trajectories are available, BRTS chooses the one most aligned with the student's current behavior; when unconditioned teacher samples fail on harder prompts, it invokes a ground-truth-conditioned recovery step to elicit a natural derivation. The selected trajectory is then used to provide reliable teacher-context supervision inside the OPD loop, augmented with an auxiliary loss on the teacher trajectory. Experiments on AIME 2024, AIME 2025, and AMC 2023 show that BRTS improves over standard OPD on challenging reasoning benchmarks, with the largest gains on harder datasets. Our code is available at https://github.com/BWGZK-keke/BRTS.

Monocular depth estimation (MDE) with self-supervised training approaches struggles in low-texture areas, where photometric losses may lead to ambiguous depth predictions. To address this, we propose a novel technique that enhances spatial information by applying a distance transform over pre-semantic contours, augmenting discriminative power in low texture regions. Our approach jointly estimates pre-semantic contours, depth and ego-motion. The pre-semantic contours are leveraged to produce new input images, with variance augmented by the distance transform in uniform areas. This approach results in more effective loss functions, enhancing the training process for depth and ego-motion. We demonstrate theoretically that the distance transform is the optimal variance-augmenting technique in this context. Through extensive experiments on KITTI, Cityscapes, Waymo, NYUv2 and ScanNet our model demonstrates robust performance, surpassing competing self-supervised methods in MDE.

3D semantic scene understanding is essential for digital twins, autonomous driving, smart agriculture, and embodied perception, yet dense point-wise annotation for point clouds remains expensive and difficult to scale. Existing annotation-free methods often face a trade-off between semantic recognition and structural efficiency: open-vocabulary and foundation-model-driven methods provide strong semantic priors, but often come with substantial computational costs, while structure-oriented methods based on superpoints, clustering, and graph reasoning are lightweight but often produce category-agnostic regions. We propose DDS, a resource-efficient structure-oriented framework for region-consistent and semanticized annotation-free 3D scene understanding. DDS preserves the lightweight superpoint-based organization paradigm while incorporating visual semantic cues from projected features and segmentation-derived masks. It first performs multi-granularity distillation to guide the 3D backbone at the point, mask-prototype, and inter-prototype levels, then applies graph diffusion over superpoints to propagate semantic information directly in 3D, producing coherent region representations without costly spectral decomposition or dense open-vocabulary 3D feature fields. Finally, DDS uses segmentation-cluster association to assign interpretable semantic names to category-agnostic 3D clusters. Experiments on real-world datasets show that DDS achieves the best performance among representative structure-oriented annotation-free baselines, improving oAcc, mAcc, and mIoU by up to 5.9%, 8.1%, and 2.4%, respectively. These results demonstrate that DDS improves region consistency and lightweight semantic recognition, providing a scalable and interpretable solution for annotation-free 3D scene understanding.

Diffusion-based models decompose sampling into many small Gaussian denoising steps -- an assumption that breaks down when generation is compressed to a few coarse transitions. Existing few-step methods address this through distillation, consistency training, or adversarial objectives, but sacrifice the likelihood framework in the process. We introduce Normalizing Trajectory Models (NTM), which models each reverse step as an expressive conditional normalizing flow with exact likelihood training. Architecturally, NTM combines shallow invertible blocks within each step with a deep parallel predictor across the trajectory, forming an end-to-end network trainable from scratch or initializable from pretrained flow-matching models. Its exact trajectory likelihood further enables self-distillation: a lightweight denoiser trained on the model's own score produces high-quality samples in four steps. On text-to-image benchmarks, NTM matches or outperforms strong image generation baselines in just four sampling steps while uniquely retaining exact likelihood over the generative trajectory.

We present PicoEyes, a unified gaze estimation framework that directly predicts all key attributes of gaze, including 3D eye parameters, eye-region segmentation, optical axis, visual axis, and depth maps, from either monocular or binocular inputs. The framework simultaneously addresses calibration, gaze forecasting, and varying device postures, while also supporting 3D eye reconstruction via joint estimation of eye parameters and depth maps in an end-to-end manner. In addition, we introduce a large-scale multi-view near-eye dataset containing comprehensive 2D and 3D annotations under diverse conditions, including train, test, rewear-test, and calibration sessions. Extensive experiments demonstrate that PicoEyes achieves state-ofthe-art performance, consistently outperforming both academic and industrial gaze tracking methods across nocalibration, calibration, rewear-after-calibration, and forecasting settings. This work establishes a practical, end-toend paradigm for robust and generalizable gaze estimation in mixed reality (MR) applications.

We introduce Ilov3Splat, a novel framework for instance-level open-vocabulary 3D scene understanding built on 3D Gaussian Splatting (3D-GS). Most prior work depends on 2D rendering-based matching or point-level semantic association, which undermines cross-view consistency, lacks coherent instance-level reasoning, and limits precision in downstream 3D tasks. To address these limitations, our method jointly optimizes scene geometry and semantic representations by augmenting Gaussian splats with view-consistent feature fields. Specifically, we leverage multi-resolution hash embedding to efficiently encode language-aligned CLIP features, enabling dense and coherent language grounding in 3D space. We further train an instance feature field using contrastive loss over SAM masks, supporting fine-grained object distinction across views. At inference time, CLIP-encoded queries are matched against the learned features, followed by two-stage 3D clustering to retrieve relevant Gaussian groups. This enables our framework to identify arbitrary objects in 3D scenes based on natural language descriptions, without requiring category supervision or manual annotations. Experiments on standard benchmarks demonstrate that Ilov3Splat outperforms prior open-vocabulary 3D-GS methods in both object selection and instance segmentation, offering a flexible and accurate solution for language-driven 3D scene understanding. Project page: https://csiro-robotics.github.io/Ilov3Splat.

In recent years, Multimodal Large Language Models (MLLMs) have achieved remarkable progress on a wide range of multimodal benchmarks. Despite these advances, most existing benchmarks mainly focus on single-image or multi-image comprehension. In real-world scenarios such as document reading, information is often presented as interleaved multimodel contexts. This requires MLLMs not only to recognize the content of individual images, but also to identify relevant textual and visual evidence, establish fine-grained alignments between them, and reason over these aligned signals in interleaved contexts based on contextual evidence. However, there is still a lack of systematic benchmarks for quantifying the fine-grained understanding ability of MLLMs in interleaved image-text contexts. To fill this gap, we propose COHERENCE, a benchmark designed to evaluate the ability of MLLMs to recover fine-grained image-text correspondences in interleaved multimodal contexts. COHERENCE covers interleaved image-text content from four representative domains and contains 6,161 high-quality questions. Moreover, we perform a six-type error analysis, enabling fine-grained attribution of failures in interleaved image-text understanding to the specific capabilities missing in current MLLMs.

Vision-language models (VLMs) have shown a promising ability in image geolocation, but they still lack structured geographic reasoning and the capacity for autonomous self-evolution. Existing methods predominantly rely on implicit parametric memory, which often exploits outdated knowledge and generates hallucinated reasoning. Furthermore, current inference is a "one-off" process, lacking the feedback loops necessary for self-evolution based on reasoning outcomes. To address these issues, we propose GeoSkill, a training-free framework based on an evolving Skill-Graph. We first initialize the graph by refining human expert trajectories into atomic, natural-language skills. For execution, GeoSkill employs an inference model to perform direct reasoning guided by the current Skill-Graph. For continuous growth, an Autonomous Evolution mechanism leverages a larger model to conduct multiple reasoning rollouts on image-coordinate pairs sourced from web-scale data and verified real-world reasoning. By analyzing both successful and failed trajectories from these rollouts, the mechanism iteratively synthesizes and prunes skills, effectively expanding the Skill-Graph and correcting geographic biases without any parameter updates. Experiments demonstrate that GeoSkill achieves promising performance in both geolocation accuracy and reasoning faithfulness on GeoRC, while maintaining superior generalization across diverse external datasets. Furthermore, our autonomous evolution fosters the emergence of novel, verifiable skills, significantly enhancing the system's cognition of real-world geographic knowledge beyond isolated case studies.

Interpreting individual neurons in deep neural networks is a crucial step towards understanding their complex decision-making processes and ensuring AI safety. Despite recent progress in neuron labeling, existing methods often limit the search space to predefined concept vocabularies or produce overly specific descriptions that fail to capture higher-order, global concepts. We introduce LINE, a novel, training-free iterative approach tailored for open-vocabulary concept labeling in vision models. Operating in a strictly black-box setting, LINE leverages a large language model and a text-to-image generator to iteratively propose and refine concepts in a closed loop, guided by activation history. LINE achieves state-of-the-art performance across multiple model architectures, yielding AUC improvements of up to 0.11 on ImageNet and 0.05 on Places365, while discovering, on average, 27% of new concepts missed by predefined vocabularies. Beyond identifying the top concept, LINE provides a complete generation history, enabling polysemanticity evaluation and producing visual explanations that rival gradient-dependent activation maximization methods. The source code will be made available soon.

Automated fact-checking is a crucial task that supports a responsible information ecosystem. While recent research has progressed from text-only to multimodal fact-checking, a prevailing assumption is that incorporating visual evidence universally improves performance. In this work, we challenge this assumption and show that the indiscriminate use of multimodal evidence can reduce accuracy. To address this challenge, we propose AMuFC, a multimodal fact-checking framework that employs two collaborative vision-language models with distinct roles for the adaptive use of visual evidence: an Analyzer determines whether visual evidence is necessary for claim verification, and a Verifier predicts claim veracity conditioned on both the retrieved evidence and the Analyzer's assessment. Experimental results on three datasets show that incorporating the Analyzer's assessment of visual evidence necessity into the Verifier's prediction yields substantial improvements in verification performance. We will release all code and datasets at https://github.com/ssu-humane/AMuFC.

Real-time depth reconstruction from ultra-high-resolution UAV imagery is essential for time-critical geospatial tasks such as disaster response, yet remains challenging due to wide-baseline parallax, large image sizes, low-texture or specular surfaces, occlusions, and strict computational constraints. Recent zero-shot diffusion models offer fast per-image dense predictions without task-specific retraining, and require fewer labelled datasets than transformer-based predictors while avoiding the rigid capture geometry requirement of classical multi-view stereo. However, their probabilistic inference prevents reliable metric accuracy and temporal consistency across sequential frames and overlapping tiles. We present ZeD-MAP, a cluster-level framework that converts a test-time diffusion depth model into a metrically consistent, SLAM-like mapping pipeline by integrating incremental cluster-based bundle adjustment (BA). Streamed UAV frames are grouped into overlapping clusters; periodic BA produces metrically consistent poses and sparse 3D tie-points, which are reprojected into selected frames and used as metric guidance for diffusion-based depth estimation. Validation on ground-marker flights captured at approximately 50 m altitude (GSD is approximately 0.85 cm/px, corresponding to 2,650 square meters ground coverage per frame) with the DLR Modular Aerial Camera System (MACS) shows that our method achieves sub-meter accuracy, with approximately 0.87 m error in the horizontal (XY) plane and 0.12 m in the vertical (Z) direction, while maintaining per-image runtimes between 1.47 and 4.91 seconds. Results are subject to minor noise from manual point-cloud annotation. These findings show that BA-based metric guidance provides consistency comparable to classical photogrammetric methods while significantly accelerating processing, enabling real-time 3D map generation.

This work presents a robust multi-class classification framework for handwritten digits that combines diffusion-driven feature denoising with a hybrid feature representation. Inspired by our previous work on brain tumor classification, the proposed approach operates in a feature space to improve the robustness to noise and adversarial attacks. This manuscript is submitted as an extended abstract rather than a full-length press-ready paper. First, the input images are converted into tight, interpretable exemplification using Non-negative Matrix Factorization (NNMF). In parallel, special deep features are extracted using a computational neural network (CNN). These integral features are combined into a united hybrid representation. The main objective of this work is to extend our previously validated two-class framework to a multi-class handwritten digit classification scenario. To improve robustness, a step diffusion operation is used in the feature space by gradually adding Gaussian noise. A feature denoiser network is trained to reverse this operation and rebuild clean representations from tilted inputs. The courteous features are then applied for multi-class classification. The suggested method is evaluated in both baseline and adversarial settings using AutoAttack. The experimental outcome present that the diffusion-based hybrid model is both effective and robust, the CNN baseline models outperforming while maintain powerful classification performance. These results explain the activity of feature-level diffusion defense for reliable multi-class handwritten digit classification.

How do multimodal models solve visual spatial tasks -- through genuine planning, or through brute-force search in token space? We introduce \textsc{MazeBench}, a benchmark of 110 procedurally generated maze images across nine controlled groups, and evaluate 16 model configurations from OpenAI, Anthropic, Google, and Alibaba. GPT-5.4 solves 91\% and Gemini 3.1 Pro 79\%, but these scores are misleading: models typically translate images into text grids and then enumerate paths step by step, consuming 1,710--22,818 tokens per solve for a task humans do quickly. Without added reasoning budgets, all configurations score only 2--12\%; on 20$\times$20 ultra-hard mazes, they hit token limits and fail. Qualitative traces reveal a common two-stage strategy: image-to-grid translation followed by token-level search, effectively BFS in prose. A text-grid ablation shows Claude Sonnet 4.6 rising from 6\% on images to 80\% when given the correct grid, isolating weak visual extraction from downstream search. When explicitly instructed not to construct a grid or perform graph search, models still revert to the same enumeration strategy. \textsc{MazeBench} therefore shows that high accuracy on visual planning tasks does not imply human-like spatial understanding.

Diffusion models achieve strong performance in generative modeling, but their success often relies heavily on classifier-free guidance (CFG), an inference-time heuristic that modifies the sampling trajectory. In theory, diffusion models trained with standard denoising score matching (DSM) should recover the target data distribution, raising two fundamental questions: (i) why is inference-time guidance necessary in practice, and (ii) can its underlying effect be internalized into a principled training objective? In this work, we argue that a key limitation of standard DSM is insufficient inter-class separation. To address this issue, we propose MCLR, an alignment objective that explicitly maximizes inter-class likelihood-ratios during training. Fine-tuning diffusion models with MCLR induces CFG-like improvements under standard sampling, substantially improving guidance-free conditional generation and narrowing the gap to inference-time CFG. Beyond these empirical benefits, we show theoretically that the CFG-guided score is exactly the optimal solution to a sample-adaptive weighted MCLR objective. This result connects CFG to alignment-based objectives, providing a mechanistic interpretation of CFG as an implicit inference-time contrastive alignment procedure.

Topology is critical in tubular structures such as blood vessels, nerve fibers, and road networks, where connectivity and loop structure govern downstream functional analysis. Vision-Language Models (VLMs) are promising candidates for understanding such structures, given their reasoning and grounding capabilities. To probe their topological perception, we systematically evaluate leading closed- and open-source VLMs on localizing and classifying four canonical topological anomalies (broken/spurious connections, missing/extra branches) in tubular-network segmentation masks. They perform nearly at random, indicating that topology-aware perception is largely absent from current general-purpose VLMs. As no existing resource pairs segmentation masks with localized anomaly annotations, we build an automated, multi-domain data-curation pipeline that synthesizes diverse topological perturbations with verifiable Betti-number annotations across graduated difficulty levels, yielding the first systematic benchmark with a large-scale training set and held-out in-distribution (ID) and out-of-distribution (OOD) test suites. Building on this benchmark, we introduce Topo-R1, centered on a topology-aware composite reward that jointly scores localization, classification, and skeleton-level structural fidelity. Supervised fine-tuning cold-starts schema-compliant outputs, and Group Relative Policy Optimization (GRPO) then optimizes the policy against this reward, steering predictions toward topologically meaningful structure rather than superficial pixel overlap. Extensive experiments show that Topo-R1 substantially outperforms general-purpose VLMs and matches or exceeds supervised baselines across ID, OOD, and real-segmentation-output protocols, establishing a strong foundation for VLM-based topological understanding of structured visual data.

Dynamic Data selection aims to accelerate training by prioritizing informative samples during online training. However, existing methods typically rely on task-specific handcrafted metrics or static/snapshot-based criteria to estimate sample importance, limiting scalability across learning paradigms and making it difficult to capture the evolving utility of data throughout training. To address this challenge, we propose Data Agent, an end-to-end dynamic data selection framework that formulates data selection as a training-aware sequential decision-making problem. The agent learns a sample-wise selection policy that co-evolves with model optimization, guided by a composite reward that integrates loss-based difficulty and confidence-based uncertainty signals. The reward signals capture complementary objectives of optimization impact and information gain, together with a tuning-free adaptive weighting mechanism that balances these signals over training. Extensive experiments across a wide range of datasets and architectures demonstrate that Data Agent consistently accelerates training while preserving or improving performance, e.g., reducing costs by over 50\% on ImageNet-1k and MMLU with lossless performance. Moreover, its dataset-agnostic formulation and modular reward make it plug-and-play across tasks and scenarios, e.g., robustness to noisy datasets, highlighting its potential in real-world scenarios. Code is available at https://github.com/Jackbrocp/Data-Agent.

The issue of algorithmic biases in deep learning has led to the development of various debiasing techniques, many of which perform complex training procedures or dataset manipulation. However, an intriguing question arises: is it possible to extract fair and bias-agnostic subnetworks from standard vanilla-trained models without relying on additional data, such as unbiased training set? In this work, we introduce Bias-Invariant Subnetwork Extraction (BISE), a learning strategy that identifies and isolates "bias-free" subnetworks that already exist within conventionally trained models, without retraining or finetuning the original parameters. Our approach demonstrates that such subnetworks can be extracted via pruning and can operate without modification, effectively relying less on biased features and maintaining robust performance. Our findings contribute towards efficient bias mitigation through structural adaptation of pre-trained neural networks via parameter removal, as opposed to costly strategies that are either data-centric or involve (re)training all model parameters. Extensive experiments on common benchmarks show the advantages of our approach in terms of the performance and computational efficiency of the resulting debiased model.

Flow matching (FM) learns vector fields by regressing stochastic velocity targets along intermediate distributions $p_t$. We identify a geometric optimization bottleneck in this regression problem: when the covariance $Σ_t$ of $p_t$ is ill-conditioned, gradient-based training rapidly fits high-variance directions while making slow progress along low-variance ones. In an exactly solvable Gaussian setting, we prove that the excess risk is weighted by $Σ_t$, and that both gradient descent and stochastic gradient descent inherit condition-number-dependent convergence. We then extend the analysis to Gaussian mixtures, showing that multimodality does not average away this effect; instead, the slowest and worst-conditioned component can control optimization. Motivated by this analysis, we propose \emph{preconditioned flow matching}, a precondition-then-match framework that transforms the target distribution into a more isotropic representation, trains the main flow in the transformed space, and maps generated samples back through the inverse transformation. We show theoretically that preconditioning reshapes the intermediate FM path and improves its conditioning. Across controlled Gaussian and Gaussian-mixture experiments, latent MNIST and other high resolution image datasets up to $512{\times}512$ resolution, preconditioning improves path-conditioning diagnostics, low-eigenvalue recovery, FID, MMD, precision, and recall. Compute-matched baselines and preconditioner-quality ablations further show that the gains are not explained merely by additional preconditioner parameters, but by improved geometry of the downstream flow matching problem.

Instruction-based video editing has witnessed rapid progress, yet current methods often struggle with precise visual control, as natural language is inherently limited in describing complex visual nuances. Although reference-guided editing offers a robust solution, its potential is currently bottlenecked by the scarcity of high-quality paired training data. To bridge this gap, we introduce a scalable data generation pipeline that transforms existing video editing pairs into high-fidelity training quadruplets, leveraging image generative models to create synthesized reference scaffolds. Using this pipeline, we construct RefVIE, a large-scale dataset tailored for instruction-reference-following tasks, and establish RefVIE-Bench for comprehensive evaluation. Furthermore, we propose a unified editing architecture, Kiwi-Edit, that synergizes learnable queries and latent visual features for reference semantic guidance. Our model achieves significant gains in instruction following and reference fidelity via a progressive multi-stage training curriculum. Extensive experiments demonstrate that our data and architecture establish a new state-of-the-art in controllable video editing. All datasets, models, and code is released at https://github.com/showlab/Kiwi-Edit.

Detecting visual anomalies in industrial inspection often requires training with only a few normal images per category. Recent few-shot methods achieve strong results employing foundation-model features, but typically rely on memory banks, auxiliary datasets, or multi-modal tuning of vision-language models. We therefore question whether such complexity is necessary given the feature representations of vision foundation models. To answer this question, we introduce SubspaceAD, a training-free method, that operates in two simple stages. First, patch-level features are extracted from a small set of normal images by a frozen DINOv2 backbone. Second, a Principal Component Analysis (PCA) model is fit to these features to estimate the low-dimensional subspace of normal variations. At inference, anomalies are detected via the reconstruction residual with respect to this subspace, producing interpretable and statistically grounded anomaly scores. Despite its simplicity, SubspaceAD achieves state-of-the-art performance across one-shot and few-shot settings without training, prompt tuning, or memory banks. In the one-shot anomaly detection setting, SubspaceAD achieves image-level and pixel-level AUROC of 97.1% and 97.5% on the MVTec-AD dataset, and 93.2% and 98.2% on the VisA dataset, respectively, surpassing prior state-of-the-art results. Code and demo are available at https://github.com/CLendering/SubspaceAD.

We investigate the feasibility of using Multimodal Large Language Models (MLLMs) for real-time online episodic memory question answering. While cloud offloading is common, it raises privacy and latency concerns for wearable assistants, hence we investigate implementation on the edge. We integrated streaming constraints into our question answering pipeline, which is structured into two asynchronous threads: a Descriptor Thread that continuously converts video into a lightweight textual memory, and a Question Answering (QA) Thread that reasons over the textual memory to answer queries. Experiments on the QAEgo4D-Closed benchmark analyze the performance of Multimodal Large Language Models (MLLMs) within strict resource boundaries, showing promising results also when compared to clound-based solutions. Specifically, an end-to-end configuration running on a consumer-grade 8GB GPU achieves 51.76% accuracy with a Time-To-First-Token (TTFT) of 0.41s. Scaling to a local enterprise-grade server yields 54.40% accuracy with a TTFT of 0.88s. In comparison, a cloud-based solution obtains an accuracy of 56.00%. These competitive results highlight the potential of edge-based solutions for privacy-preserving episodic memory retrieval.

Test-time training (TTT) with KV binding as sequence modeling layer is commonly interpreted as a form of online meta-learning that memorizes a key-value mapping at test time. However, our analysis reveals multiple phenomena that contradict this memorization-based interpretation. Motivated by these findings, we revisit the formulation of TTT and show that a broad class of TTT architectures can be expressed as a form of learned linear attention operator. Beyond explaining previously puzzling model behaviors, this perspective yields multiple practical benefits: it enables principled architectural simplifications, admits fully parallel formulations that preserve performance while improving efficiency, and provides a systematic reduction of diverse TTT variants to a standard linear attention form. Overall, our results reframe TTT not as test-time memorization, but as learned linear attention with enhanced representational capacity. Project page: https://research.nvidia.com/labs/sil/projects/tttla/.

Estimating simulation-ready scenes from real-world observations is crucial for downstream planning and policy learning tasks. Regretfully, existing methods struggle in cluttered environments, often exhibiting prohibitive computational cost, poor robustness, and restricted generality when scaling to multiple interacting objects. We propose a unified optimization-based formulation for real-to-sim scene estimation that jointly recovers the shapes and poses of multiple rigid objects under physical constraints. Our method is built on two key technical innovations. First, we leverage the recently introduced shape-differentiable contact model, whose global differentiability permits joint optimization over object geometry and pose while modeling inter-object contacts. Second, we exploit the structured sparsity of the augmented Lagrangian Hessian to derive an efficient linear system solver whose computational cost scales favorably with scene complexity. Building on this formulation, we develop an end-to-end Simulation-ready Physics-Aware Reconstruction for Cluttered Scenes (SPARCS) pipeline, which integrates learning-based object initialization, physics-constrained joint shape-pose optimization, and differentiable texture refinement. Experiments on cluttered scenes with up to 5 objects and 22 convex hulls demonstrate that our approach robustly reconstructs physically valid, simulation-ready object shapes and poses. Project webpage: https://rory-weicheng.github.io/SPARCS/.

Despite the remarkable success of sampling-based generative models such as flow matching, they can still produce samples of inconsistent or degraded quality. To assess sample reliability and generate higher-quality outputs, we propose uncertainty-aware flow matching (UA-Flow), a lightweight extension of flow matching that predicts the velocity field together with heteroscedastic uncertainty. UA-Flow estimates per-sample uncertainty by propagating velocity uncertainty through the flow dynamics. These uncertainty estimates act as a reliability signal for individual samples, and we further use them to steer generation via uncertainty-aware classifier guidance and classifier-free guidance. Experiments on image generation show that UA-Flow produces uncertainty signals more highly correlated with sample fidelity than baseline methods, and that uncertainty-guided sampling further improves generation quality.

Agents in cyber-physical systems are increasingly entrusted with safety-critical tasks. Ensuring safety of these agents often requires localizing the pose for subsequent actions. Pose estimates can, e.g., be obtained from various combinations of lidar sensors, cameras, and external services such as GPS. Crucially, in safety-critical domains, a rough estimate is insufficient to formally determine safety, i.e., guaranteeing safety even in the worst-case scenario, and external services might additionally not be trustworthy. We address this problem by presenting a certified pose estimation in 3D solely from a camera image and a well-known target geometry. This is realized by formally bounding the pose, which is computed by leveraging recent results from reachability analysis and formal neural network verification. Our experiments demonstrate that our approach efficiently and accurately localizes agents in both synthetic and real-world experiments.

Vision-language models such as CLIP often struggle to faithfully understand long, detail-rich captions, relying on dominant scene cues while overlooking fine-grained visual evidence. We propose a hierarchical vision-language learning principle for understanding scenes as part-to-whole compositions: before forming a whole-scene representation, a model should uncover what semantic parts appear where in the image. To this end, we propose CAFT (Cross-domain Alignment of Forests and Trees), a vision-language model that jointly learns local text-region alignment at intermediate representations and global image-text alignment at the final representation. Exploiting the organization of long captions, where local descriptions often correspond to scene parts, CAFT employs a fine-to-coarse image encoder and a part-whole text encoder to discover localized part semantics and progressively compose them into a global image-text representation. Trained on 30M image-text pairs, CAFT achieves state-of-the-art performance on six long-text retrieval benchmarks and exhibits strong scaling behavior. Experiments show that CAFT learns fine-grained representations that localize textual semantics in image regions without explicit region-level supervision.