视觉大模型 / VLM - arXivDaily 专题

2512.04599 2025-12-05 cs.CV 70%

Malicious Image Analysis via Vision-Language Segmentation Fusion: Detection, Element, and Location in One-shot

通过视觉-语言分割融合进行恶意图像分析：一次检测、元素识别与定位

Sheng Hang, Chaoxiang He, Hongsheng Hu, Hanqing Hu, Bin Benjamin Zhu, Shi-Feng Sun, Dawu Gu, Shuo Wang

机构 * Shanghai Jiao Tong University（上海交通大学）； Microsoft Corporation（微软公司）

专题命中视觉定位与Grounding ：vision-language model(abstract)；VLM(abstract)；分类 cs.CV

AI总结本文提出一种零样本方法，通过视觉-语言分割融合实现恶意图像的检测、元素识别与定位，提升细粒度审核的准确性和鲁棒性。

2511.22826 2025-12-04 cs.CV 70%

Some Modalities are More Equal Than Others: Decoding and Architecting Multimodal Integration in MLLMs

某些模态比其他模态更平等：在MLLMs中解码和架构多模态整合

Tianle Chen, Chaitanya Chakka, Arjun Reddy Akula, Xavier Thomas, Deepti Ghadiyaram

机构 * Boston University（波士顿大学）； Google DeepMind（谷歌DeepMind）

专题命中视觉定位与Grounding ：grounding(abstract)；multimodal large language model(abstract)；分类 cs.CV

AI总结本文研究了多模态大语言模型对矛盾模态的鲁棒性，提出模态对齐调优策略以提升多模态推理能力。

2511.01755 2025-12-02 cs.CV cs.RO 70%

3EED: Ground Everything Everywhere in 3D

3EED: 在三维中万物皆 grounded

Rong Li, Yuhao Dong, Tianshuai Hu, Ao Liang, Youquan Liu, Dongyue Lu, Liang Pan, Lingdong Kong, Junwei Liang, Ziwei Liu

机构 * WorldBench Team（WorldBench团队）

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

AI总结 3EED提出一个大规模多平台多模态三维 grounding 基准测试，通过提供丰富的户外场景数据和跨平台学习技术，推动语言驱动的三维具身感知研究。

Comments NeurIPS 2025 DB Track; 38 pages, 17 figures, 10 tables; Project Page at https://project-3eed.github.io/

2510.23594 2025-12-02 cs.CV 70%

PRISM-Bench: A Benchmark of Puzzle-Based Visual Tasks with CoT Error Detection

PRISM-Bench: 一个包含推理错误检测的基于谜题的视觉任务基准

Yusu Qian, Cheng Wan, Chao Jia, Yinfei Yang, Qingyu Zhao, Zhe Gan

专题命中视觉定位与Grounding ：visual reasoning(abstract)；multimodal large language model(abstract)；分类 cs.CV

AI总结 PRISM-Bench通过检测推理错误评估多模态模型的视觉推理能力，揭示了流畅生成与忠实推理之间的差距。

Comments This paper's first error detection task's ground truth data contains hallucination introduced by gpt and needs to be withdrawn

2511.11165 2025-12-02 cs.CV 70%

Explainable Deep Convolutional Multi-Type Anomaly Detection

可解释的深度卷积多类型异常检测

Alex George, Lyudmila Mihaylova, Sean Anderson

机构 * School of Electrical and Electronic Engineering（电子工程学院）

专题命中视觉定位与Grounding ：vision-language model(abstract)；VLM(abstract)；分类 cs.CV

AI总结本文提出 MultiTypeFCDD 框架，通过图像级别标签生成多通道热图，实现多类型异常检测，有效解决传统方法的计算和内存限制问题。

2511.22170 2025-12-01 cs.CV 70%

Partially Shared Concept Bottleneck Models

部分共享概念瓶颈模型

Delong Zhao, Qiang Huang, Di Yan, Yiqun Sun, Jun Yu

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

AI总结 PS-CBM通过部分共享的概念策略和概念效率准确性度量，提升模型的分类准确性和可解释性。

Comments 14 pages, 7 figures, 11 tables, Accepted to AAAI 2026

2511.21984 2025-12-01 cs.CV 70%

PPBoost: Progressive Prompt Boosting for Text-Driven Medical Image Segmentation

PPBoost: 逐步提示增强用于文本驱动的医学图像分割

Xuchen Li, Hengrui Gu, Mohan Zhang, Qin Liu, Zhen Tan, Xinyuan Zhu, Huixue Zhou, Tianlong Chen, Kaixiong Zhou

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

AI总结 PPBoost通过逐步增强弱文本提示为强空间指导，提升医学图像分割的精度与鲁棒性。

2511.21705 2025-12-01 cs.CL cs.CV 70%

Insight-A: Attribution-aware for Multimodal Misinformation Detection

Insight-A: 多模态虚假信息检测中的归因意识

Junjie Wu, Yumeng Fu, Chen Gong, Guohong Fu

机构 * School of Computer Science and Technology, Soochow University（苏州大学计算机科学与技术学院）； Institute of Artificial Intelligence, Soochow University（苏州大学人工智能研究院）； School of Computer Science and Technology, Harbin Institute of Technology（哈尔滨工业大学计算机科学与技术学院）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.CV

AI总结 Insight-A通过归因意识和分层推理提升多模态虚假信息检测效果，有效识别伪造来源并增强跨模态一致性检查。

2511.19111 2025-11-27 cs.CV 70%

DiffSeg30k: A Multi-Turn Diffusion Editing Benchmark for Localized AIGC Detection

DiffSeg30k: 一种用于局部AIGC检测的多轮扩散编辑基准

Hai Ci, Ziheng Peng, Pei Yang, Yingxin Xuan, Mike Zheng Shou

机构 * Show Lab, National University of Singapore（新加坡国立大学Show实验室）； South China University of Technology（华南理工大学）

专题命中视觉定位与Grounding ：vision-language model(abstract)；VLM(abstract)；分类 cs.CV

AI总结 DiffSeg30k通过引入多轮扩散编辑基准，推动AIGC检测从二元分类到语义分割的转变，提升局部编辑定位和模型识别的可靠性。

Comments 16 pages, 10 figures; typos corrected, references added

2511.19920 2025-11-26 cs.CV 70%

Intelligent Image Search Algorithms Fusing Visual Large Models

融合视觉大模型的智能图像搜索算法

Kehan Wang, Tingqiong Cui, Yang Zhang, Yu Chen, Shifeng Wu, Zhenzhang Li

机构 * Chongqing University（重庆大学）； CRRC Chongqing Co., Ltd.（CRRC重庆公司）； Guangdong Polytechnic Normal University（广东 polytechnic 正规大学）

专题命中视觉定位与Grounding ：VLM(abstract)；grounding(abstract)；分类 cs.CV

AI总结 DetVLM融合视觉大模型与物体检测，实现细粒度图像检索中的状态搜索和零样本搜索，取得高准确率。

Comments 31 pages,7 figures

2511.15165 2025-11-25 cs.CR cs.AI 70%

Can MLLMs Detect Phishing? A Comprehensive Security Benchmark Suite Focusing on Dynamic Threats and Multimodal Evaluation in Academic Environments

MLLMs能否检测钓鱼？一个全面的安全基准套件，聚焦于动态威胁和学术环境中的多模态评估

Jingzhuo Zhou

机构 * School of Computer Science and Engineering（计算机科学与工程学院）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.AI

AI总结本文提出AdapT-Bench，一个针对学术环境动态钓鱼攻击的多模态安全基准套件，旨在评估MLLM的防御能力。

2511.12363 2025-11-18 cs.CV 70%

Explainable AI-Generated Image Detection RewardBench

Michael Yang, Shijian Deng, William T. Doan, Kai Wang, Tianyu Yang, Harsh Singh, Yapeng Tian

机构 * The University of Texas at Dallas（德克萨斯大学达拉斯分校）； University of Toronto（多伦多大学）； University of Notre Dame（诺特大学）； Stony Brook University（石溪大学）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.CV

2511.09955 2025-11-14 cs.CV 70%

Robust Object Detection with Pseudo Labels from VLMs using Per-Object Co-teaching

Uday Bhaskar, Rishabh Bhattacharya, Avinash Patel, Sarthak Khoche, Praveen Anil Kulkarni, Naresh Manwani

机构 * Machine Learning Lab IIIT Hyderabad（IIIT Hyderabad 机器学习实验室）； Bosch Global Software Technologies（博世全球软件技术公司）

专题命中视觉定位与Grounding ：vision-language model(abstract)；VLM(abstract)；分类 cs.CV

2511.08971 2025-11-13 cs.HC cs.CV cs.MM 70%

Plug-and-Play Clarifier: A Zero-Shot Multimodal Framework for Egocentric Intent Disambiguation

Sicheng Yang, Yukai Huang, Weitong Cai, Shitong Sun, You He, Jiankang Deng, Hang Zhang, Jifei Song, Zhensong Zhang

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

Comments 16 pages, 9 figures, AAAI 2026

2511.07744 2025-11-12 cs.CV 70%

VectorSynth: Fine-Grained Satellite Image Synthesis with Structured Semantics

Daniel Cher, Brian Wei, Srikumar Sastry, Nathan Jacobs

机构 * Washington University in St. Louis（华盛顿大学圣路易斯分校）

专题命中视觉定位与Grounding ：vision language model(abstract)；grounding(abstract)；分类 cs.CV

2509.24378 2025-11-12 cs.LG 70%

AXIS: Explainable Time Series Anomaly Detection with Large Language Models

Tian Lan, Hao Duong Le, Jinbo Li, Wenjun He, Meng Wang, Chenghao Liu, Chen Zhang

机构 * Huawei（华为）

专题命中视觉定位与Grounding ：vision language model(abstract)；grounding(abstract)；分类 cs.LG

2507.14686 2025-11-12 cs.CV 70%

From Semantics, Scene to Instance-awareness: Distilling Foundation Model for Grounded Open-vocabulary Situation Recognition

Chen Cai, Tianyi Liu, Jianjun Gao, Wenyang Liu, Kejun Wu, Ruoyu Wang, Yi Wang, Soo Chin Liew

机构 * National University of Singapore（新加坡国立大学）； Nanyang Technological University（南洋理工大学）； Huazhong University of Science and Technology（华中科技大学）； The Hong Kong Polytechnic University（香港理工大学）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.CV

2511.06281 2025-11-11 cs.CV 70%

VideoSSR: Video Self-Supervised Reinforcement Learning

Zefeng He, Xiaoye Qu, Yafu Li, Siyuan Huang, Daizong Liu, Yu Cheng

机构 * Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）； The Chinese University of Hong Kong（香港中文大学）； Shanghai Jiao Tong University（上海交通大学）； Wuhan University（武汉大学）

专题命中视觉定位与Grounding ：grounding(abstract)；multimodal large language model(abstract)；分类 cs.CV

2406.19875 2025-11-11 cs.CV 70%

InfiniBench: A Benchmark for Large Multi-Modal Models in Long-Form Movies and TV Shows

Kirolos Ataallah, Eslam Abdelrahman, Mahmoud Ahmed, Chenhui Gou, Khushbu Pahwa, Jian Ding, Mohamed Elhoseiny

机构 * KAUST（卡塔尔科技大学）； Monash University（墨尔本大学）； RICE University（里士满大学）

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

Comments Accepted for oral presentation at the EMNLP 2025 main conference

2505.16495 2025-11-06 cs.CV 70%

ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation

Lingfeng Wang, Hualing Lin, Senda Chen, Tao Wang, Changxu Cheng, Yangyang Zhong, Dong Zheng, Wuyue Zhao

机构 * Uni-Ubi ； Zhejiang University（浙江大学）； Tongji University（同济大学）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.CV

2511.00940 2025-11-04 cs.RO cs.AI 70%

URDF-Anything: Constructing Articulated Objects with 3D Multimodal Language Model

Zhe Li, Xiang Bai, Jieyu Zhang, Zhuangzhe Wu, Che Xu, Ying Li, Chengkai Hou, Shanghang Zhang

机构 * State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University（多媒体信息处理国家重点实验室，计算机学院，北京大学）； University of Washington（华盛顿大学）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.AI

Comments Accepted to the 39th Conference on Neural Information Processing Systems (NeurIPS 2025)

2511.00613 2025-11-04 cs.CV 70%

CueBench: Advancing Unified Understanding of Context-Aware Video Anomalies in Real-World

Yating Yu, Congqi Cao, Zhaoying Wang, Weihua Meng, Jie Li, Yuxin Li, Zihao Wei, Zhongpei Shen, Jiajun Zhang

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

2510.26006 2025-10-31 cs.CV cs.CL 70%

CAVE: Detecting and Explaining Commonsense Anomalies in Visual Environments

Rishika Bhagwatkar, Syrielle Montariol, Angelika Romanou, Beatriz Borges, Irina Rish, Antoine Bosselut

机构 * EPFL（瑞士联邦理工学院）； MILA（蒙特利尔人工智能研究院）

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

Journal ref 2025 Conference on Empirical Methods in Natural Language Processing

2510.23473 2025-10-28 cs.CV 70%

Video-Thinker: Sparking "Thinking with Videos" via Reinforcement Learning

Shijian Wang, Jiarui Jin, Xingjian Wang, Linxin Song, Runhao Fu, Hecheng Wang, Zongyuan Ge, Yuan Lu, Xuelian Cheng

机构 * Southeast University（东南大学）； Monash University（墨尔本大学）； Xiaohongshu Inc.（小红书公司）； University of Southern California（南加州大学）； Fudan University（复旦大学）

专题命中视觉定位与Grounding ：grounding(abstract)；multimodal large language model(abstract)；分类 cs.CV

2510.23190 2025-10-28 cs.CV 70%

Evaluation of Vision-LLMs in Surveillance Video

Pascal Benschop, Cristian Meo, Justin Dauwels, Jelte P. Mense

机构 * Department of Computer Science（计算机科学系）； Delft University of Technology（代尔夫特理工大学）； LatentWorlds AI ； National Policelab AI & Model-Driven Decisions Lab（国家警务实验室AI与模型驱动决策实验室）

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

Comments Accepted as poster in the NeurIPS 2025 Workshop on Space in Vision, Language, and Embodied AI

2510.20803 2025-10-24 cs.CV 70%

ARGenSeg: Image Segmentation with Autoregressive Image Generation Model

Xiaolong Wang, Lixiang Ru, Ziyuan Huang, Kaixiang Ji, Dandan Zheng, Jingdong Chen, Jun Zhou

机构 * Ant Group（蚂蚁集团）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.CV

Comments Accepted to NeurIPS 2025, 18 pages

2510.16785 2025-10-21 cs.CV 70%

Segmentation as A Plug-and-Play Capability for Frozen Multimodal LLMs

Jiazhen Liu, Long Chen

机构 * Department of Computer Science and Engineering（计算机科学与工程系）

专题命中视觉定位与Grounding ：multimodal large language model(abstract)；MLLM(abstract)；分类 cs.CV

2510.14583 2025-10-17 cs.CV cs.CL 70%

Talking Points: Describing and Localizing Pixels

Matan Rusanovsky, Shimon Malnick, Shai Avidan

机构 * Tel Aviv University（特拉维夫大学）

专题命中视觉定位与Grounding ：vision-language model(abstract)；grounding(abstract)；分类 cs.CV

2510.12798 2025-10-15 cs.CV 70%

Detect Anything via Next Point Prediction

Qing Jiang, Junan Huo, Xingyu Chen, Yuda Xiong, Zhaoyang Zeng, Yihao Chen, Tianhe Ren, Junzhi Yu, Lei Zhang

机构 * International Digital Economy Academy (IDEA)（国际数字经济学院（IDEA））

专题命中视觉定位与Grounding ：grounding(abstract)；MLLM(abstract)；分类 cs.CV

Comments homepage: https://rex-omni.github.io/

2510.12693 2025-10-15 cs.AI 70%

ERA: Transforming VLMs into Embodied Agents via Embodied Prior Learning and Online Reinforcement Learning

Hanyang Chen, Mark Zhao, Rui Yang, Qinwei Ma, Ke Yang, Jiarui Yao, Kangrui Wang, Hao Bai, Zhenhailong Wang, Rui Pan, Mengchao Zhang, Jose Barreiros, Aykut Onol, ChengXiang Zhai, Heng Ji, Manling Li, Huan Zhang, Tong Zhang

机构 * UIUC（伊利诺伊大学香槟分校）； Northwestern University（西北大学）； Toyota Research Institute（丰田研究院）

专题命中视觉定位与Grounding ：vision language model(abstract)；grounding(abstract)；分类 cs.AI

1. 视觉定位与Grounding 7370 篇

Malicious Image Analysis via Vision-Language Segmentation Fusion: Detection, Element, and Location in One-shot

Some Modalities are More Equal Than Others: Decoding and Architecting Multimodal Integration in MLLMs

3EED: Ground Everything Everywhere in 3D

PRISM-Bench: A Benchmark of Puzzle-Based Visual Tasks with CoT Error Detection

Explainable Deep Convolutional Multi-Type Anomaly Detection

Partially Shared Concept Bottleneck Models

PPBoost: Progressive Prompt Boosting for Text-Driven Medical Image Segmentation

Insight-A: Attribution-aware for Multimodal Misinformation Detection

DiffSeg30k: A Multi-Turn Diffusion Editing Benchmark for Localized AIGC Detection

Intelligent Image Search Algorithms Fusing Visual Large Models

Can MLLMs Detect Phishing? A Comprehensive Security Benchmark Suite Focusing on Dynamic Threats and Multimodal Evaluation in Academic Environments

Explainable AI-Generated Image Detection RewardBench

Robust Object Detection with Pseudo Labels from VLMs using Per-Object Co-teaching

Plug-and-Play Clarifier: A Zero-Shot Multimodal Framework for Egocentric Intent Disambiguation

VectorSynth: Fine-Grained Satellite Image Synthesis with Structured Semantics

AXIS: Explainable Time Series Anomaly Detection with Large Language Models

From Semantics, Scene to Instance-awareness: Distilling Foundation Model for Grounded Open-vocabulary Situation Recognition

VideoSSR: Video Self-Supervised Reinforcement Learning

InfiniBench: A Benchmark for Large Multi-Modal Models in Long-Form Movies and TV Shows

ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation

URDF-Anything: Constructing Articulated Objects with 3D Multimodal Language Model

CueBench: Advancing Unified Understanding of Context-Aware Video Anomalies in Real-World

CAVE: Detecting and Explaining Commonsense Anomalies in Visual Environments

Video-Thinker: Sparking "Thinking with Videos" via Reinforcement Learning

Evaluation of Vision-LLMs in Surveillance Video

ARGenSeg: Image Segmentation with Autoregressive Image Generation Model

Segmentation as A Plug-and-Play Capability for Frozen Multimodal LLMs

Talking Points: Describing and Localizing Pixels

Detect Anything via Next Point Prediction

ERA: Transforming VLMs into Embodied Agents via Embodied Prior Learning and Online Reinforcement Learning