3D 视觉 - arXivDaily 专题

2508.19204 2025-08-27 cs.CV cs.AI cs.GR 73%

LSD-3D: Large-Scale 3D Driving Scene Generation with Geometry Grounding

Julian Ost, Andrea Ramazzina, Amogh Joshi, Maximilian Bömer, Mario Bijelic, Felix Heide

机构 * Julian Ost（未知）； Andrea Ramazzina（未知）； Amogh Joshi（未知）； Maximilian Bömer（未知）； Mario Bijelic（未知）； Felix Heide（未知）

专题命中新视角合成：novel view synthesis(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

Comments Project webpage: https://light.princeton.edu/LSD-3D

2504.18768 2025-05-02 cs.GR cs.CV 73%

TransparentGS: Fast Inverse Rendering of Transparent Objects with Gaussians

Letian Huang, Dongwei Ye, Jialin Dan, Chengzhi Tao, Huiwen Liu, Kun Zhou, Bo Ren, Yuanqi Li, Yanwen Guo, Jie Guo

机构 * State Key Lab for Novel Software Technology, Nanjing University（南京大学新型软件技术国家重点实验室）； TMCC, College of Computer Science, Nankai University（天津大学计算机学院）； State Key Lab of CAD & CG, Zhejiang University（浙江大学计算机辅助设计与图形学国家重点实验室）； Institute of Hangzhou Holographic Intelligent Technology（杭州全息智能技术研究院）

专题命中新视角合成：Gaussian Splatting(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments accepted by SIGGRAPH 2025; https://letianhuang.github.io/transparentgs/

2504.15776 2025-04-28 cs.CV cs.RO 73%

Pose Optimization for Autonomous Driving Datasets using Neural Rendering Models

Quentin Herau, Nathan Piasco, Moussab Bennehar, Luis Roldão, Dzmitry Tsishkou, Bingbing Liu, Cyrille Migniot, Pascal Vasseur, Cédric Demonceaux

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.RO

Comments under review

2412.05256 2025-03-14 cs.CV cs.AI cs.LG cs.RO 73%

Extrapolated Urban View Synthesis Benchmark

Xiangyu Han, Zhen Jia, Boyi Li, Yan Wang, Boris Ivanovic, Yurong You, Lingjie Liu, Yue Wang, Marco Pavone, Chen Feng, Yiming Li

专题命中新视角合成：Gaussian Splatting(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.RO

Comments Project page: https://ai4ce.github.io/EUVS-Benchmark/

2503.03751 2025-03-06 cs.CV cs.GR 73%

GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

Xuanchi Ren, Tianchang Shen, Jiahui Huang, Huan Ling, Yifan Lu, Merlin Nimier-David, Thomas Müller, Alexander Keller, Sanja Fidler, Jun Gao

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments To appear in CVPR 2025. Website: https://research.nvidia.com/labs/toronto-ai/GEN3C/

2407.00500 2024-07-02 cs.CV cs.AI cs.GR cs.LG 73%

Intrinsic PAPR for Point-level 3D Scene Albedo and Shading Editing

Alireza Moazeni, Shichong Peng, Ke Li

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

2405.03659 2024-06-12 cs.CV cs.GR 73%

A Construct-Optimize Approach to Sparse View Synthesis without Camera Pose

Kaiwen Jiang, Yang Fu, Mukund Varma T, Yash Belhe, Xiaolong Wang, Hao Su, Ravi Ramamoorthi

专题命中新视角合成：Gaussian Splatting(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

2310.17994 2024-04-25 cs.CV cs.GR 73%

ZeroNVS: Zero-Shot 360-Degree View Synthesis from a Single Image

Kyle Sargent, Zizhang Li, Tanmay Shah, Charles Herrmann, Hong-Xing Yu, Yunzhi Zhang, Eric Ryan Chan, Dmitry Lagun, Li Fei-Fei, Deqing Sun, Jiajun Wu

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments Accepted to CVPR 2024. 12 pages

2311.11825 2024-04-09 cs.CV cs.GR 73%

Holistic Inverse Rendering of Complex Facade via Aerial 3D Scanning

Zixuan Xie, Rengan Xie, Rong Li, Kai Huang, Pengju Qiao, Jingsen Zhu, Xu Yin, Qi Ye, Wei Hua, Yuchi Huo, Hujun Bao

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

2402.14415 2024-02-23 cs.CV cs.GR 73%

TaylorGrid: Towards Fast and High-Quality Implicit Field Learning via Direct Taylor-based Grid Optimization

Renyi Mao, Qingshan Xu, Peng Zheng, Ye Wang, Tieru Wu, Rui Ma

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

2402.03445 2024-02-22 cs.CV cs.GR cs.LG 73%

Denoising Diffusion via Image-Based Rendering

Titas Anciukevičius, Fabian Manhardt, Federico Tombari, Paul Henderson

专题命中新视角合成：3D reconstruction(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments Accepted at ICLR 2024. Project page: https://anciukevicius.github.io/generative-image-based-rendering

2210.07181 2023-06-06 cs.CV cs.LG cs.RO 73%

MonoNeRF: Learning Generalizable NeRFs from Monocular Videos without Camera Pose

Yang Fu, Ishan Misra, Xiaolong Wang

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.RO

Comments ICML 2023 camera ready version. Project page: https://oasisyang.github.io/mononerf

2304.13386 2023-06-05 cs.CV cs.AI cs.GR 73%

VGOS: Voxel Grid Optimization for View Synthesis from Sparse Inputs

Jiakai Sun, Zhanjie Zhang, Jiafu Chen, Guangyuan Li, Boyan Ji, Lei Zhao, Wei Xing, Huaizhong Lin

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments IJCAI 2023 Accepted (Main Track)

2106.10689 2023-02-02 cs.CV cs.GR 73%

NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction

Peng Wang, Lingjie Liu, Yuan Liu, Christian Theobalt, Taku Komura, Wenping Wang

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments 23 pages

2011.14398 2022-08-02 cs.CV cs.GR 73%

RGBD-Net: Predicting color and depth images for novel views synthesis

Phong Nguyen-Ha, Animesh Karnewar, Lam Huynh, Esa Rahtu, Jiri Matas, Janne Heikkila

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments 19 pages, 15 figures. Code will be available at: https://github.com/phongnhhn92/RGBDNet

2207.04465 2022-07-12 cs.CV cs.GR 73%

Progressively-connected Light Field Network for Efficient View Synthesis

Peng Wang, Yuan Liu, Guying Lin, Jiatao Gu, Lingjie Liu, Taku Komura, Wenping Wang

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments Project page: https://totoro97.github.io/projects/prolif

2112.01473 2022-06-08 cs.CV cs.GR 73%

Neural Point Light Fields

Julian Ost, Issam Laradji, Alejandro Newell, Yuval Bahat, Felix Heide

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments 9 pages, replacement changed font of equations

2205.15768 2022-06-01 cs.CV cs.GR cs.LG 73%

SAMURAI: Shape And Material from Unconstrained Real-world Arbitrary Image collections

Mark Boss, Andreas Engelhardt, Abhishek Kar, Yuanzhen Li, Deqing Sun, Jonathan T. Barron, Hendrik P. A. Lensch, Varun Jampani

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

2204.10477 2022-05-05 cs.CV cs.GR 73%

Learning Dynamic View Synthesis With Few RGBD Cameras

Shengze Wang, YoungJoong Kwon, Yuan Shen, Qian Zhang, Andrei State, Jia-Bin Huang, Henry Fuchs

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments One of the coauthors believes that this work should be improved more before releasing it on arXiv, and thus suggested withdrawing this paper. There will not be a replacement for this paper

2112.15399 2022-04-12 cs.CV cs.GR eess.IV 73%

InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Rendering

Mijeong Kim, Seonguk Seo, Bohyung Han

专题命中新视角合成：NeRF(abstract)；novel view synthesis(abstract)；分类 cs.CV、cs.GR

Comments CVPR 2022, Website: http://cv.snu.ac.kr/research/InfoNeRF

2607.19765 2026-07-23 cs.CV 新提交 70%

Extending a Large View Synthesis Model for Multi-view Panoptic Segmentation

扩展用于多视图全景分割的大视图合成模型

Kwonyoung Ryu, In-Jae Lee, Jonghyun Jin, Hyunjee Lee, Jongmin Lee, Jaesik Park

机构 * POSTECH（浦项科技大学）； POSCO DX（浦项制铁DX公司）； Chung-Ang University（中央大学）； Seoul National University（首尔国立大学）

专题命中新视角合成：3D reconstruction(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结研究将大视图合成模型从外观渲染扩展到3D场景理解，提出复用冻结模型传播全景标签的分割管道，无需3D重建和特定训练，在ScanNet和Replica数据集上实验，分割质量佳且新视图合成表现优。

Comments Accepted to ECCV 2026. 31 pages, 7 figures, 9 tables. Project page: https://kwonyoung9120.github.io/PanopticLVSM/

2607.19731 2026-07-23 cs.GR cs.ET cs.MM eess.IV eess.SP 新提交 70%

Fast Wave-optics Rendering of Multiplane Images for 3D Holographic Displays

用于3D全息显示的多平面图像快速波光学渲染

Brian Chao, Dario Seyb, Nathan Matsuda, Oliver Cossairt, Yang Zhou, Douglas Lanman, Gordon Wetzstein, Grace Kuo, Changwon Jang

专题命中新视角合成：3D reconstruction(abstract)；novel view synthesis(abstract)；分类 cs.GR

AI总结针对神经渲染在2D设备上查看限制感知真实感问题，提出基于多平面图像的波光学渲染管道，该算法在运行时大幅超越现有算法，速度提升显著，且图像质量优，经多数据集验证性能出色。

2607.02301 2026-07-03 cs.CV 新提交 70%

InvSplat: Inverse Feed-Forward Scene Splatting

InvSplat: 逆前馈场景高斯泼溅

Polina Karpikova, Wenjing Bian, Haofei Xu, Hendrik Lensch, Andreas Geiger

专题命中新视角合成：3D reconstruction(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结提出一种前馈多视图重建框架，直接预测具有材质属性的结构化3D高斯表示，实现几何与反射率的联合估计，支持重光照和新视角合成。

2605.29891 2026-05-29 cs.CV 70%

DVSM: Decoder-only View Synthesis Model Done Right

DVSM: 正确的仅解码器视图合成模型

Cheng Sun, Jaesung Choe, Min-Hung Chen, Ryo Hachiuma, Yu-Chiang Frank Wang

机构 * NVIDIA ； National Taiwan University（国立台湾大学）

专题命中新视角合成：3DGS(abstract,abstract_cn)；分类 cs.CV

AI总结提出仅解码器架构DVSM，通过隐式KV-cache表示场景，在相同渲染复杂度下以更少参数超越编码器-解码器变体，并利用共享权重、基础模型先验和分阶段块大小优化效率与质量，在多个基准上实现新视点合成的最优结果。

Comments Code at https://github.com/NVLabs/dvsm

2512.14180 2026-05-26 cs.CV 70%

Spherical Voronoi: Directional Appearance as a Differentiable Partition of the Sphere

球面Voronoi：作为球面可微分划分的定向外观

Francesco Di Sario, Daniel Rebain, Dor Verbin, Marco Grangetto, Andrea Tagliasacchi

机构 * University of Torino（都灵大学）； Simon Fraser University（西蒙弗雷泽大学）； University of British Columbia（不列颠哥伦比亚大学）； University of Toronto（多伦多大学）； Google DeepMind（谷歌DeepMind）

专题命中新视角合成：Gaussian Splatting(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结提出球面Voronoi（SV）作为3D高斯泼溅中外观表示的统一框架，通过可学习区域划分实现视图依赖效果，在反射建模上达到最先进水平。

2506.05317 2026-05-20 cs.CV 70%

ProJo4D: Progressive Joint Optimization for Sparse-View Inverse Physics Estimation

ProJo4D：渐进式联合优化用于稀疏视图逆物理估计

Daniel Rho, Jun Myeong Choi, Biswadip Dey, Roni Sengupta

机构 * University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）； Meta Reality Labs（Meta现实实验室）

专题命中新视角合成：3D reconstruction(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结本文提出ProJo4D，一种渐进式联合优化框架，用于解决稀疏视图下逆物理参数估计问题，通过逐步扩展联合优化参数集，提高了4D未来状态预测和物理参数估计的准确性，达到几何精度提升10倍的性能。

Comments TMLR 2026

2507.01110 2026-04-30 cs.GR cs.LG 70%

A LoD of Gaussians: Unified Training and Rendering for Ultra-Large Scale Reconstruction with External Memory

高斯点云的层次结构：用于超大规模重建的统一训练与渲染，借助外部内存

Felix Windisch, Thomas Köhler, Lukas Radl, Mattia D'Urso, Michael Steiner, Dieter Schmalstieg, Markus Steinberger

机构 * Graz University of Technology（格拉茨技术大学）； University of Stuttgart（斯图加特大学）； Huawei Technologies（华为技术）

专题命中新视角合成：Gaussian Splatting(abstract)；novel view synthesis(abstract)；分类 cs.GR

AI总结本文提出一种无需分块的高斯点云框架，通过外存存储完整场景并动态流式传输相关高斯点，实现超大规模场景的实时训练与渲染，支持多尺度重建与交互式可视化。

Journal ref Proceedings of SIGGRAPH 2026

2604.07250 2026-04-09 cs.CV 70%

Geo-EVS: Geometry-Conditioned Extrapolative View Synthesis for Autonomous Driving

Geo-EVS：基于几何条件的外推视图合成用于自动驾驶

Yatong Lan, Rongkui Tang, Lei He

机构 * School of Vehicle and Mobility, Tsinghua University（清华大学车辆与运载学院）； State Key Laboratory of Intelligent Green Vehicle and Mobility, Tsinghua University（清华大学智能绿色车辆与移动国家重点实验室）； School of Science, Minzu University（中央民族大学理学院）； Chongqing Changan Automobile Co., Ltd.（重庆长安汽车股份有限公司）

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结 Geo-EVS通过几何条件化框架提升自动驾驶中稀疏视图合成的几何精度，尤其在高角度和低覆盖率场景中表现优异，并改进下游3D检测性能。

2603.29734 2026-04-01 cs.CV 70%

GRVS: a Generalizable and Recurrent Approach to Monocular Dynamic View Synthesis

GRVS：一种通用且递归的方法用于单目动态视图合成

Thomas Tanay, Mohammed Brahimi, Michal Nazarczuk, Qingwen Zhang, Sibi Catley-Chandar, Arthur Moreau, Zhensong Zhang, Eduardo Pérez-Pellitero

专题命中新视角合成：Gaussian Splatting(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结本文提出GRVS方法，通过递归循环和动态输入平面扫描，实现高精度的单目动态视图合成，优于现有多种基于高斯点划和扩散的方法。

Comments CVPR Findings 2026

2603.15433 2026-03-17 cs.CV 70%

Real-Time Human Frontal View Synthesis from a Single Image

从单张图像实时合成人体正面视图

Fangyu Lin, Yingdong Hu, Lunjie Zhu, Zhening Liu, Yushi Huang, Zehong Lin, Jun Zhang

机构 * HKUST（香港科技大学）

专题命中新视角合成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV

AI总结本文提出PrismMirror框架，通过级联学习策略实现单图像实时正面视图合成，提升视觉真实性和结构准确性，达到24 FPS的实时性能。

1. 新视角合成 733 篇

LSD-3D: Large-Scale 3D Driving Scene Generation with Geometry Grounding

TransparentGS: Fast Inverse Rendering of Transparent Objects with Gaussians

Pose Optimization for Autonomous Driving Datasets using Neural Rendering Models

Extrapolated Urban View Synthesis Benchmark

GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

Intrinsic PAPR for Point-level 3D Scene Albedo and Shading Editing

A Construct-Optimize Approach to Sparse View Synthesis without Camera Pose

ZeroNVS: Zero-Shot 360-Degree View Synthesis from a Single Image

Holistic Inverse Rendering of Complex Facade via Aerial 3D Scanning

TaylorGrid: Towards Fast and High-Quality Implicit Field Learning via Direct Taylor-based Grid Optimization

Denoising Diffusion via Image-Based Rendering

MonoNeRF: Learning Generalizable NeRFs from Monocular Videos without Camera Pose

VGOS: Voxel Grid Optimization for View Synthesis from Sparse Inputs

NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction

RGBD-Net: Predicting color and depth images for novel views synthesis

Progressively-connected Light Field Network for Efficient View Synthesis

Neural Point Light Fields

SAMURAI: Shape And Material from Unconstrained Real-world Arbitrary Image collections

Learning Dynamic View Synthesis With Few RGBD Cameras

InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Rendering

Extending a Large View Synthesis Model for Multi-view Panoptic Segmentation

Fast Wave-optics Rendering of Multiplane Images for 3D Holographic Displays

InvSplat: Inverse Feed-Forward Scene Splatting

DVSM: Decoder-only View Synthesis Model Done Right

Spherical Voronoi: Directional Appearance as a Differentiable Partition of the Sphere

ProJo4D: Progressive Joint Optimization for Sparse-View Inverse Physics Estimation

A LoD of Gaussians: Unified Training and Rendering for Ultra-Large Scale Reconstruction with External Memory

Geo-EVS: Geometry-Conditioned Extrapolative View Synthesis for Autonomous Driving

GRVS: a Generalizable and Recurrent Approach to Monocular Dynamic View Synthesis

Real-Time Human Frontal View Synthesis from a Single Image