CVPR 2026 Oral Papers Collection

A community-maintained index of 141 CVPR 2026 Oral papers, organized by research topic, with direct links to the virtual conference page, arXiv preprints and code repositories where available.

中文版本 / Chinese README

Status

• Total orals: 141
• arXiv links: 102 (72%)
• GitHub links: 29 (20%)

Papers not yet on arXiv may be uploaded closer to the conference — PRs welcome.

Per-category counts

Sub-categories may cross-list a paper (e.g. 3D editing appears in both §1 and §2); the table de-duplicates within each top-level category, so the column sum exceeds 141.

Files

Contributing

If you spot a missing arXiv or GitHub link, feel free to open an issue or submit a pull request.

---

1. 3D Reconstruction, Geometry, Gaussian / Radiance Field, SfM, Registration

1.1 3D Reconstruction, Gaussian Splatting, NeRF, Dynamic Scenes

• 40261 GLINT: Modeling Scene-Scale Transparency via Gaussian Radiance Transport — Youngju Na et al. · arxiv
• 40277 Energy-GS: Image Energy-guided Pose Alignment Gaussian Splatting with redesigned pose gradient flow — Yu Gao et al.
• 40279 SmokeSVD: Smoke Reconstruction from A Single View via Progressive Novel View Synthesis and Refinement with Diffusion Models — Chen Li et al. · arxiv
• 40287 GaussianFluent: Gaussian Simulation for Dynamic Scenes with Mixed Materials — Bei Huang et al. · arxiv
• 40289 Proxy-GS: Unified Occlusion Priors for Training and Inference in Structured 3D Gaussian Splatting — Yuanyuan Gao et al. · arxiv
• 40291 Z-Order Transformer for Feed-Forward Gaussian Splatting — Can Wang et al.
• 40312 RetimeGS: Continuous-Time Reconstruction of 4D Gaussian Splatting — Xuezhen Wang et al. · arxiv
• 40313 PhyGaP: Physically-Grounded Gaussians with Polarization Cues — Jiale Wu et al. · arxiv
• 40333 MeshSplatting: Differentiable Rendering with Opaque Meshes — Jan Held et al. · arxiv
• 40345 Efficiently Reconstructing Dynamic Scenes one D4RT at a Time — Chuhan Zhang et al. · arxiv
• 40355 Selfi: Self Improving Reconstruction Engine via 3D Geometric Feature Alignment — Youming Deng et al. · arxiv
• 40358 4D Primitive-Mâché: Glueing Primitives for Persistent 4D Scene Reconstruction — Kirill Mazur et al.
• 40372 Evidential Neural Radiance Fields — Ruxiao Duan et al. · arxiv code
• 40381 VGGT-$\Omega$ — Jianyuan Wang et al.
• 40395 Chorus: Multi-Teacher Pretraining for Holistic 3D Gaussian Scene Encoding — Yue Li et al. · arxiv

1.2 Reflection, Transparency, Materials, Lighting, Inverse Rendering

• 40274 PPISP: Physically-Plausible Compensation and Control of Photometric Variations in Radiance Field Reconstruction — Isaac Deutsch et al. · arxiv code
• 40290 POLAR: A Portrait OLAT Dataset and Generative Framework for Illumination-Aware Face Modeling — Zhuo Chen et al. · arxiv
• 40310 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture Objects — Zhicheng Liang et al.
• 40314 Dual Band Video Thermography: Separating Time-Varying Reflection and Emission Near Ambient Conditions — Sriram Narayanan et al.
• 40328 OLATverse: A Large-scale Real-world Object Dataset with Precise Lighting Control — Xilong Zhou et al. · arxiv
• 40356 Towards Photorealistic and Efficient Bokeh Rendering via Diffusion Framework — Linxiao Shi et al.
• 40359 SeeGroup: Multi-Layer Depth Estimation of Transparent Surfaces via Self-Determined Grouping — Hongyu Wen et al.
• 40390 Relightable Holoported Characters: Capturing and Relighting Dynamic Human Performance from Sparse Views — Kunwar Maheep Singh et al. · arxiv
• 40393 UnReflectAnything: RGB-Only Highlight Removal by Rendering Synthetic Specular Supervision — Alberto Rota et al. · arxiv

1.3 Multi-view Geometry, SfM, Cameras, Pose Graph

• 40264 QuadSync: Quadrifocal Tensor Synchronization via Tucker Decomposition — Daniel Miao et al. · arxiv code
• 40265 Breaking the Scalability Limit of Multi-Projector Calibration with Embedded Cameras — Takumi Kawano et al.
• 40268 PoseGAM: Robust Unseen Object Pose Estimation via Geometry-Aware Multi-View Reasoning — Jianqi Chen et al. · arxiv
• 40275 From Pairs to Sequences: Track-Aware Policy Gradients for Keypoint Detection — yepeng liu et al. · arxiv
• 40302 Global-Aware Edge Prioritization for Pose Graph Initialization — Tong Wei et al. · arxiv code
• 40320 TESO: Online Tracking of Essential Matrix by Stochastic Optimization — Jaroslav Moravec et al. · arxiv
• 40341 Linear Fundamental Matrix Estimation from 7 or 5 Points — Taci Ata Kucukpinar et al.

1.4 3D Representations, Shape Decomposition, Shape Modeling

• 40259 Learning Convex Decomposition via Feature Fields — Yuezhi Yang et al. · arxiv
• 40276 Learning Eigenstructures of Unstructured Data Manifolds — Roy Velich et al. · arxiv
• 40283 Native and Compact Structured Latents for 3D Generation — Jianfeng XIANG et al. · arxiv
• 40293 Residual Primitive Fitting of 3D Shapes with SuperFrusta — Aditya Ganeshan et al. · arxiv
• 40323 FILTR: Extracting Topological Features from Pretrained 3D Models — Louis Martinez et al.
• 40325 FAITHFUL CONTOURING: NEAR-LOSSLESS 3D VOXEL REPRESENTATION FREE FROM ISO-SURFACE — Yihao Luo et al. · arxiv
• 40370 SAM 3D: 3Dfy Anything in Images — Xingyu Chen et al. · arxiv
• 40380 AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned Flows — Zhenglin Zhou et al. · arxiv code

1.5 Occupancy / Scene Completion / Active Mapping

• 40286 SparseWorld-TC: Trajectory-Conditioned Sparse Occupancy World Model — Jiayuan Du et al. · arxiv
• 40308 OccuFly: A 3D Vision Benchmark for Semantic Scene Completion from the Aerial Perspective — Markus Gross et al. · arxiv code
• 40324 Monocular Open Vocabulary Occupancy Prediction for Indoor Scenes — Changqing Zhou et al. · arxiv code
• 40384 MAGICIAN: Efficient Long-Term Planning with Imagined Gaussians for Active Mapping — Shiyao Li et al. · arxiv

1.6 Multi-view / Point Cloud Registration

• 40330 FUSER: Feed-Forward Multiview 3D Registration Transformer and SE(3)$^N$ Diffusion Refinement — Haobo Jiang et al. · arxiv
• 40391 ComPose: A Unified Completion-Pose Framework for Robust Category-Level Object Pose Estimation — Huan Ren et al.

2. Generative Models, Diffusion, Image / Video / 3D Generation & Editing

2.1 Image Generation & Editing

• 40271 A Style is Worth One Code: Unlocking Code-to-Style Image Generation with Discrete Style Space — Huijie Liu et al. · arxiv code
• 40301 SliderEdit: Continuous Image Editing with Fine-Grained Instruction Control — Arman Zarei et al. · arxiv
• 40304 PixelDiT: Pixel Diffusion Transformers for Image Generation — Yongsheng Yu et al. · arxiv code
• 40306 Guiding a Diffusion Model by Swapping Its Tokens — Weijia Zhang et al. · arxiv
• 40319 Breaking Semantic Boundaries: Distribution-Guided Semantic Exploration for Creative Generation — Fu Feng et al.
• 40352 ChordEdit: One-Step Low-Energy Transport for Image Editing — Liangsi Lu et al. · arxiv
• 40353 SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models — Jiwoo Chung et al. · arxiv
• 40360 SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware Caching — Yasaman Haghighi et al. · arxiv

2.2 Video Generation & Related

• 40256 ProPhy: Progressive Physical Alignment for Dynamic World Simulation — Zijun Wang et al. · arxiv
• 40300 Thinking with Drafts: Speculative Temporal Reasoning for Efficient Long-Video Understanding — Pengfei Hu et al. · arxiv
• 40338 Texvent: Asynchronous Event Data Simulation via Text Prompt — Ruofei Wang et al.
• 40357 TEAR: Temporal-aware Automated Red-teaming for Text-to-Video Models — Jiaming He et al. · arxiv
• 40364 PAVAS: Physics-Aware Video-to-Audio Synthesis — Oh Hyun-Bin et al. · arxiv

2.3 3D Generation & Editing

• 40283 Native and Compact Structured Latents for 3D Generation — Jianfeng XIANG et al. · arxiv
• 40344 SPARK: Sim-ready Part-level Articulated Reconstruction with VLM Knowledge — Yumeng He et al. · arxiv
• 40368 3D-LATTE: Latent Space 3D Editing from Textual Instructions — Maria Parelli et al. · arxiv
• 40380 AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned Flows — Zhenglin Zhou et al. · arxiv code

2.4 Multimodal Generation / Tokenizer / Interleaved

• 40278 Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding — Christopher Clark et al. · arxiv
• 40282 AToken: A Unified Tokenizer for Vision — Jiasen Lu et al. · arxiv
• 40336 Weaver: Decoupled Training for Interleaved Multi-modal Generation — Jinbo Xing et al.

3. Medical Imaging: VLM, Segmentation, Reconstruction, Registration

3.1 Medical Segmentation

• 40258 SegMoTE: Token-Level Mixture of Experts for Medical Image Segmentation — Yujie Lu et al. · arxiv
• 40281 DK-DDIL: Adaptive Knowledge Retention for Dynamic Domain-Incremental Learning in Medical Imaging — Yuxi Ma et al.
• 40317 Dual-level Adapter Boosting Prompt-free Curvilinear Structure Segmentation — Kai Zhu et al.
• 40334 R$^2$-Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and Statistical Rejection — Shuaike Shen et al. · arxiv code
• 40347 MDCS-MoAME: Multi-directional Composite Scanning with Mixture of Attention and Mamba Experts for Cancer Survival Prediction — Linjie Qu et al.
• 40369 Medic-AD: : Towards Medical Vision-Language Model's Clinical Intelligence — Woohyeon Park et al.
• 40373 CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report Generation — Pablo Messina et al. · arxiv code

3.2 Medical Reconstruction / Inverse Problems / Imaging

• 40257 Neural Field-Based 3D Surface Reconstruction of Microstructures from Multi-Detector Signals in Scanning Electron Microscopy — Shuo Chen et al. · arxiv code
• 40269 Efficient unrolled networks for large-scale 3D inverse problems — Romain Vo et al. · arxiv
• 40298 Differentiable Vector Quantization for Rate-Distortion Optimization of Generative Image Compression — SHIYIN JIANG et al. · arxiv code
• 40307 Hearing the Room Through the Shape of the Drum: Modal-Guided Sound Recovery from Multi-Point Surface Vibrations — Shai Bagon et al.
• 40309 Spectrum from Defocus: Fast Spectral Imaging with Chromatic Focal Stack — M. Kerem Aydin et al. · arxiv
• 40361 MetaSpectra+: A Compact Broadband Metasurface Camera for Snapshot Hyperspectral+ Imaging — Yuxuan Liu et al. · arxiv

3.3 Medical VLMs / Uncertainty / Reliability

• 40270 LATA: Laplacian-Assisted Transductive Adaptation for Conformal Uncertainty in Medical VLMs — Behzad Bozorgtabar et al. · arxiv
• 40369 Medic-AD: : Towards Medical Vision-Language Model's Clinical Intelligence — Woohyeon Park et al.
• 40373 CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report Generation — Pablo Messina et al. · arxiv code

3.4 Medical Registration

• 40376 Learning Diffeomorphism for Medical Image Registration with Time-Embedded Architectures Using Semigroup Regularization — Mohammadjavad Matinkia et al. · arxiv

4. Segmentation, Detection, Recognition, Tracking, Correspondence

4.1 Image / Video Segmentation

• 40263 INSID3: Training-Free In-Context Segmentation with DINOv3 — Claudia Cuttano et al. · arxiv code
• 40267 S$^2$AM3D: Scale-controllable Part Segmentation of 3D Point Clouds — Han Su et al.
• 40296 VGGT-Segmentor: Geometry-Enhanced Cross-View Segmentation — Yulu Gao et al. · arxiv
• 40297 RobotSeg: A Model and Dataset for Segmenting Robots in Image and Video — Haiyang Mei et al. · arxiv code
• 40321 PR-MaGIC: Prompt Refinement Via Mask Decoder Gradient Flow For In-Context Segmentation — Minjae Lee et al. · arxiv
• 40362 CoSMo3D: Open-World Promptable 3D Semantic Segmentation through LLM-Guided Canonical Spatial Modeling — Li Jin et al. · arxiv
• 40374 Differentiable Laplacian Matrix Guided Superpixel Segmentation — Jeremy Juybari et al.

4.2 Detection / OOD / Long-tail / Category Discovery

• 40266 Learning Latent Concepts for Detecting Out-of-Distribution Objects — Ting Peng et al.
• 40292 ANTS: Adaptive Negative Textual Space Shaping for OOD Detection via Test-Time MLLM Understanding and Reasoning — Wenjie Zhu et al. · arxiv code
• 40316 Learning Like Humans: Analogical Concept Learning for Generalized Category Discovery — Jizhou Han et al. · arxiv code
• 40340 Confusion-Aware Spectral Regularizer for Long-Tailed Recognition — Ziquan Zhu et al. · arxiv

4.3 Correspondence, Keypoints, Optical Flow, Tracking, Representations

• 40275 From Pairs to Sequences: Track-Aware Policy Gradients for Keypoint Detection — yepeng liu et al. · arxiv
• 40295 Customized Fusion: A Closed-Loop Dynamic Network for Adaptive Multi-Task-Aware Infrared-Visible Image Fusion — Zengyi Yang et al. · arxiv code
• 40305 MARCO: Navigating the Unseen Space of Semantic Correspondence — Claudia Cuttano et al. · arxiv code
• 40322 The SA-FARI Dataset: Segment Anything in Footage of Animals for Recognition and Identification — Dante Wasmuht et al. · arxiv
• 40348 Optical Flow Matching: Reframing Optical Flow as Continuous Transport Dynamics — Ao Luo et al.
• 40366 Featurising Pixels from Dynamic 3D Scenes with Linear In-Context Learners — Nikita Araslanov et al.
• 40367 Streaming Diffusion Model for Fast Infrared and Visible Video Fusion — Jinyuan Liu et al.
• 40392 SEATrack: Simple, Efficient, and Adaptive Multimodal Tracker — Junbin Su et al. · arxiv code
• 40394 SDTrack: A Baseline for Event-based Tracking via Spiking Neural Networks — Yimeng Shan et al. · arxiv
• 40396 U$^{2}$Flow: Uncertainty-Aware Unsupervised Optical Flow Estimation — Xunpei Sun et al. · arxiv code

4.4 Specialized Recognition / Human Motion / Dance / Sign Language

• 40280 ImmerIris: A Large-Scale Dataset and Benchmark for Off-Axis and Unconstrained Iris Recognition in Immersive Applications — Yuxi Mi et al. · arxiv
• 40311 SAM 3D Body: Robust Full-Body Human Mesh Recovery — Xitong Yang et al. · arxiv code
• 40331 Plant Taxonomy Meets Plant Counting: A Fine-Grained, Taxonomic Dataset for Counting Hundreds of Plant Species — Jinyu Xu et al. · arxiv code
• 40349 Fine-grained Image Aesthetic Assessment: Learning Discriminative Scores from Relative Ranks — Zhichao Yang et al. · arxiv
• 40350 MAMMA: Markerless Accurate Multi-person Motion Acquisition — Hanz Cuevas Velasquez et al.
• 40351 Natural Human Motion Recovery by Aligning High-Order Temporal Dynamics from Monocular Videos — Dingkun Wei et al.
• 40354 SoccerMaster: A Vision Foundation Model for Soccer Understanding — Haolin Yang et al. · arxiv
• 40375 BoostSLT: Boosting Sign Language Translation via a Plug-and-Play Diffusion-Based Semantic Enhancer — Changzhou Han et al.
• 40388 OpenDance: Multimodal Controllable 3D Dance Generation with Large-scale Internet Data — Jinlu Zhang et al. · arxiv

5. VLM / MLLM / Multimodal Reasoning / Safety / Benchmarks

5.1 Multimodal Reasoning & Capability Analysis

• 40284 VS-Bench: Evaluating VLMs for Strategic Abilities in Multi-Agent Environments — Zelai Xu et al. · arxiv
• 40303 GeoViS: Geospatially Rewarded Visual Search for Remote Sensing Visual Grounding — Peirong Zhang et al. · arxiv
• 40318 FINER: MLLMs Hallucinate under Fine-grained Negative Queries — Rui Xiao et al. · arxiv
• 40335 PAI-Bench: A Comprehensive Benchmark For Physical AI — Fengzhe Zhou et al. · arxiv
• 40378 Understanding Task Transfer in Vision-Language Models — Bhuvan Sachdeva et al.
• 40382 InfiniBench: Infinite Benchmarking for Visual Spatial Reasoning with Customizable Scene Complexity — Haoming Wang et al. · arxiv
• 40386 CodeV: Code with Images for Faithful Visual Reasoning via Tool-Aware Policy Optimization — Xinhai Hou et al. · arxiv

5.2 MLLM / VLM Safety & Attack / Defense

• 40299 Adversarial Style Optimization: Enhancing VLM Jailbreaks by GRPO-based Stylistic Triggers Optimization — Bingjun Luo et al.
• 40357 TEAR: Temporal-aware Automated Red-teaming for Text-to-Video Models — Jiaming He et al. · arxiv
• 40371 ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior — Weikai Lu et al. · arxiv

5.3 Video / Vision-Language Model Capabilities

• 40278 Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding — Christopher Clark et al. · arxiv
• 40326 PAS: A Training-Free Stabilizer for Temporal Encoding in Video LLMs — Bowen Sun et al. · arxiv

6. Robotics, Embodied AI, Navigation, Manipulation, Autonomous Driving, World Models

6.1 Embodied / Navigation / Manipulation

• 40327 Memory-Augmented Scene Understanding and Exploration for Open-World Aerial Object-Goal Navigation — Jiacong Zhou et al.
• 40337 SocialNav: Training Human-Inspired Foundation Model for Socially-Aware Embodied Navigation — Ziyi Chen et al. · arxiv
• 40363 AT-VLA: Adaptive Tactile Injection for Enhanced Feedback Reaction in Vision-Language-Action Models — Xiaoqi Li et al.
• 40389 Structural Action Transformer for 3D Dexterous Manipulation — Xiaohan Lei et al. · arxiv

6.2 Autonomous Driving / World Models / Scenario Mining

• 40329 WorldLens: Full-Spectrum Evaluations of Driving World Models in Real World — Alan Liang et al. · arxiv
• 40346 RefAV: Towards Planning Centric Scenario Mining — Cainan Davidson et al. · arxiv code
• 40379 Learning to Drive via Real-World Simulation at Scale — Haochen Tian et al. · arxiv code

6.3 Game Agents

• 40365 NitroGen: An Open Foundation Model for Generalist Gaming Agents — Loïc Magne et al. · arxiv

7. Model Compression, Training Efficiency, Optimization, Federated Learning, Task Arithmetic

7.1 Compression / Acceleration / Edge

• 40260 Mapping Networks — Lord Sen et al.
• 40262 NuWa: Deriving Lightweight Class-Specific Vision Transformers for Edge Devices — Ziteng Wei et al.
• 40385 Does YOLO Really Need to See Every Training Image in Every Epoch? — Xingxing Xie et al. · arxiv

7.2 Inference & Architecture Efficiency

• 40326 PAS: A Training-Free Stabilizer for Temporal Encoding in Video LLMs — Bowen Sun et al. · arxiv
• 40332 ViT$^3$: Unlocking Test-Time Training in Vision — Dongchen Han et al. · arxiv code
• 40353 SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models — Jiwoo Chung et al. · arxiv
• 40360 SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware Caching — Yasaman Haghighi et al. · arxiv

7.3 Optimization / Federated / Clustering / Theory

• 40272 FedAdamom: Adaptive Momentum for Improved Generalization in Federatedd Optimization — Wenjie Hou et al.
• 40273 Advancing Image Classification with Discrete Diffusion Classification Modeling — Omer Belhasin et al. · arxiv code
• 40288 Scalable Multi-View Subspace Clustering with Tensorized Anchor Guidance — Miao Jia et al.
• 40387 Understanding and Enforcing Weight Disentanglement in Task Arithmetic — Shangge Liu et al. · arxiv code

7.4 Dataset Distillation / Data Quality

• 40285 Rethinking Dataset Distillation: Hard Truths about Soft Labels — Priyam Dey et al. · arxiv
• 40377 Data Leakage Detection and De-duplication in Large Scale Geospatial Image Datasets — Yeshwanth Kumar Adimoolam et al. · arxiv code

8. Privacy, Security, Watermarking, Copyright, Membership Inference

• 40294 Erasing Invisible Watermarks via Novel View Synthesis — Fahad Shamshad et al. · arxiv
• 40315 Black-box Membership Inference Attacks on the Pre-training Data of Image-generation Models — Tao Qi et al.
• 40339 Revisiting Geometric Obfuscation with Dual Convergent Lines for Privacy-Preserving Image Queries in Visual Localization — Jeonggon Kim et al.
• 40342 NOWA: Null-space Optical Watermark for Invisible Capture Fingerprinting and Tamper Localization — Edwin Vargas et al. · arxiv
• 40383 LDP-Slicing: Local Differential Privacy for Images via Randomized Bit-Plane Slicing — Yuanming Cao et al. · arxiv

9. Datasets & Benchmarks

• 40280 ImmerIris: A Large-Scale Dataset and Benchmark for Off-Axis and Unconstrained Iris Recognition in Immersive Applications — Yuxi Mi et al. · arxiv
• 40308 OccuFly: A 3D Vision Benchmark for Semantic Scene Completion from the Aerial Perspective — Markus Gross et al. · arxiv code
• 40310 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture Objects — Zhicheng Liang et al.
• 40322 The SA-FARI Dataset: Segment Anything in Footage of Animals for Recognition and Identification — Dante Wasmuht et al. · arxiv
• 40328 OLATverse: A Large-scale Real-world Object Dataset with Precise Lighting Control — Xilong Zhou et al. · arxiv
• 40329 WorldLens: Full-Spectrum Evaluations of Driving World Models in Real World — Alan Liang et al. · arxiv
• 40331 Plant Taxonomy Meets Plant Counting: A Fine-Grained, Taxonomic Dataset for Counting Hundreds of Plant Species — Jinyu Xu et al. · arxiv code
• 40335 PAI-Bench: A Comprehensive Benchmark For Physical AI — Fengzhe Zhou et al. · arxiv
• 40343 CineBrain: A Large-Scale Multi-Modal Brain Dataset During Naturalistic Audiovisual Narrative Processing — Jianxiong Gao et al. · arxiv
• 40354 SoccerMaster: A Vision Foundation Model for Soccer Understanding — Haolin Yang et al. · arxiv
• 40382 InfiniBench: Infinite Benchmarking for Visual Spatial Reasoning with Customizable Scene Complexity — Haoming Wang et al. · arxiv