ICCV 2025 accepted papers on Video
416 of the 2,701 papers accepted at ICCV 2025 (15.4%) are on Video, compared with 11.5% at ICCV 2023. They fall into 4 topics.
See ICCV 2025 in the interactive map
Submitting to ICCV? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Video generation · 203 papers
- VACE: All-in-One Video Creation and Editing (343 citations)
- ReCamMaster: Camera-Controlled Generative Rendering from A Single Video (203 citations)
- OmniHuman-1: Rethinking the Scaling-Up of One-Stage Conditioned Human Animation Models (122 citations)
- TrajectoryCrafter: Redirecting Camera Trajectory for Monocular Videos via Diffusion Models (117 citations)
- WorldScore: A Unified Evaluation Benchmark for World Generation (109 citations)
- Phantom: Subject-Consistent Video Generation via Cross-Modal Alignment (103 citations)
- GameFactory: Creating New Games with Generative Interactive Videos (87 citations)
- Long Context Tuning for Video Generation (87 citations)
- MagicDrive-V2: High-Resolution Long Video Generation for Autonomous Driving with Adaptive Control (80 citations)
- Adaptive Caching for Faster Video Generation with Diffusion Transformers (73 citations)
- Epona: Autoregressive Diffusion World Model for Autonomous Driving (72 citations)
- CameraCtrl II: Dynamic Scene Exploration via Camera-controlled Video Diffusion Models (71 citations)
and 191 more in the interactive map.
Video segmentation and tracking · 95 papers
- SAM2Long: Enhancing SAM 2 for Long Video Segmentation with a Training-Free Memory Tree (73 citations)
- BVINet: Unlocking Blind Video Inpainting with Zero Annotations (57 citations)
- STAR: Spatial-Temporal Augmentation with Text-to-Video Models for Real-World Video Super-Resolution (53 citations)
- DiffVSR: Revealing an Effective Recipe for Taming Robust Video Super-Resolution Against Complex Degradations (32 citations)
- ReferDINO: Referring Video Object Segmentation with Visual Grounding Foundations (28 citations)
- AllTracker: Efficient Dense Point Tracking at High Resolution (25 citations)
- D3: Training-Free AI-Generated Video Detection Using Second-Order Features (23 citations)
- Vulnerability-Aware Spatio-Temporal Learning for Generalizable Deepfake Video Detection (21 citations)
- Low-Light Image Enhancement Using Event-Based Illumination Estimation (17 citations)
- RoMo: Robust Motion Segmentation Improves Structure from Motion (17 citations)
- Beyond Spatial Frequency: Pixel-wise Temporal Frequency-based Deepfake Video Detection (16 citations)
- CAVIS: Context-Aware Video Instance Segmentation (14 citations)
and 83 more in the interactive map.
Video-language understanding · 91 papers
- LVBench: An Extreme Long Video Understanding Benchmark (278 citations)
- Describe Anything: Detailed Localized Image and Video Captioning (80 citations)
- LVAgent: Long Video Understanding by Multi-Round Dynamical Collaboration of MLLM Agents (58 citations)
- Q-Frame: Query-aware Frame Selection and Multi-Resolution Adaptation for Video-LLMs (58 citations)
- FrameFusion: Combining Similarity and Importance for Video Token Reduction on Large Vision Language Models (49 citations)
- VideoLLaMB: Long Streaming Video Understanding with Recurrent Memory Bridges (46 citations)
- Flash-VStream: Efficient Real-Time Understanding for Long Video Streams (43 citations)
- ViLLa: Video Reasoning Segmentation with Large Language Model (40 citations)
- Vamba: Understanding Hour-Long Videos with Hybrid Mamba-Transformers (35 citations)
- InstructSeg: Unifying Instructed Visual Segmentation with Multi-modal Large Language Models (33 citations)
- OphCLIP: Hierarchical Retrieval-Augmented Learning for Ophthalmic Surgical Video-Language Pretraining (31 citations)
- MINERVA: Evaluating Complex Video Reasoning (27 citations)
and 79 more in the interactive map.
Action recognition · 27 papers
- MMAD: Multi-label Micro-Action Detection in Videos (55 citations)
- Joint Self-Supervised Video Alignment and Action Segmentation (11 citations)
- Learning to Generalize without Bias for Open-Vocabulary Action Recognition (10 citations)
- Error Recognition in Procedural Videos using Generalized Task Graph (8 citations)
- Hierarchical-aware Orthogonal Disentanglement Framework for Fine-grained Skeleton-based Action Recognition (8 citations)
- What Changed and What Could Have Changed? State-Change Counterfactuals for Procedure-Aware Video Representation Learning (8 citations)
- Multi-Modal Few-Shot Temporal Action Segmentation (7 citations)
- Open-World Skill Discovery from Unsegmented Demonstration Videos (6 citations)
- MistSense: Versatile Online Detection of Procedural and Execution Mistakes (5 citations)
- Rethinking Key-frame-based Micro-expression Recognition: A Robust and Accurate Framework Against Key-frame Errors (5 citations)
- Towards Efficient General Feature Prediction in Masked Skeleton Modeling (5 citations)
- Diffusion-Based Imaginative Coordination for Bimanual Manipulation (4 citations)
and 15 more in the interactive map.
