CVPR 2026 accepted papers on Video
587 of the 4,042 papers accepted at CVPR 2026 (14.5%) are on Video, compared with 13.6% at CVPR 2025. They fall into 4 topics.
See CVPR 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Video generation · 278 papers
- PC-Talk: Precise Facial Animation Control for Audio-Driven Talking Face Generation (1 citation)
- Taming Video Models for 3D and 4D Generation via Zero-Shot Camera Control (1 citation)
- A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens
- Accelerating Diffusion-based Video Editing via Heterogeneous Caching: Beyond Full Computing at Sampled Denoising Timestep
- ActAvatar: Temporally-Aware Precise Action Control for Talking Avatars
- ActionMesh: Animated 3D Mesh Generation with Temporal 3D Diffusion
- Active Intelligence in Video Avatars via Closed-loop World Modeling
- AdaCluster: Adaptive Query-Key Clustering for Sparse Attention in Video Generation
- Adaptive Video Distillation: Mitigating Oversaturation and Temporal Collapse in Few-Step Generation
- AlcheMinT: Fine-grained Temporal Control for Multi-Reference Consistent Video Generation
- Anchoring and Rescaling Attention for Semantically Coherent Inbetweening
- Ani3DHuman: Photorealistic 3D Human Animation with Self-guided Stochastic Sampling
and 266 more in the interactive map.
Video-language understanding · 149 papers
- HumanVBench: Probing Human-Centric Video Understanding in MLLMs with Automatically Synthesized Benchmarks (1 citation)
- Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding (1 citation)
- PAS: A Training-Free Stabilizer for Temporal Encoding in Video LLMs (1 citation)
- Skyra: AI-Generated Video Detection via Grounded Artifact Reasoning (1 citation)
- 4D-RGPT: Toward Region-level 4D Understanding via Perceptual Distillation
- A Multi-Agent Perception-Action Alliance for Efficient Long Video Reasoning
- Accelerating Streaming Video Large Language Models via Hierarchical Token Compression
- ActivityForensics: A Comprehensive Benchmark for Localizing Manipulated Activity in Videos
- AdaSpark: Adaptive Sparsity for Efficient Long-Video Understanding
- An Empirical Study on How Video-LLMs Answer Video Questions
- APPO: Attention-guided Perception Policy Optimization for Video Reasoning
- Attend Before Attention: Efficient and Scalable Video Understanding via Autoregressive Gazing
and 137 more in the interactive map.
Video segmentation and tracking · 127 papers
- FlashVSR: Towards Real-time Diffusion-Based Streaming Video Super Resolution (1 citation)
- SAGA: Source Attribution of Generative AI Videos (1 citation)
- AdaSpot: Spend Resolution Where It Matters for Precise Event Spotting
- AR2-4FV: Anchored Referring and Re-identification for Long-Term Grounding in Fixed-View Videos
- Bootstrapping Video Semantic Segmentation Model via Distillation-assisted Test-Time Adaptation
- CaptionFormer: Unified Segmentation, Tracking, and Captioning for Spatio-Temporal Objects
- CoCoVideo: The High-Quality Commercial-Model-Based Contrastive Benchmark for AI-Generated Video Detection
- Color When It Counts: Grayscale-Guided Online Triggering for Always-On Streaming Video Sensing
- Compressed-Domain-Aware Online Video Super-Resolution
- Continuous Exposure-Time Modeling for Realistic Atmospheric Turbulence Synthesis
- CoWTracker: Tracking by Warping instead of Correlation
- DeDelayed: Deleting Remote Inference Delay via On-Device Correction
and 115 more in the interactive map.
Action recognition · 33 papers
- MS-Temba: Multi-Scale Temporal Mamba for Understanding Long Untrimmed Videos (1 citation)
- A Stitch in Time: Learning Procedural Workflow via Self-Supervised Plackett-Luce Ranking
- Action Motifs: Self-Supervised Hierarchical Representation of Human Body Movements
- Back to the Feature: Explaining Video Classifiers with Video Counterfactual Explanations
- CoMo: Learning Continuous Latent Motion from Internet Videos for Scalable Robot Learning
- Decompose and Transfer: CoT-Prompting Enhanced Alignment for Open-Vocabulary Temporal Action Detection
- DETACH : Decomposed Spatio-Temporal Alignment for Exocentric Video and Ambient Sensors with Staged Learning
- Exploring Adaptive Masked Reconstruction for Self-Supervised Skeleton-Based Action Recognition
- Face-Guided Sentiment Boundary Enhancement for Weakly-Supervised Temporal Sentiment Localization
- Frame2Freq: Spectral Adapters for Fine-Grained Video Understanding
- GEM-TFL: Bridging Weak and Full Supervision for Forgery Localization through EM-Guided Decomposition and Temporal Refinement
- Hierarchical Action Learning for Weakly-Supervised Action Segmentation
and 21 more in the interactive map.