TPAMI 2025 papers on Video
74 of the 932 papers published in TPAMI 2025 (7.9%) are on Video, compared with 9.8% at TPAMI 2024. They fall into 4 topics.
See TPAMI 2025 in the interactive map
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 segmentation and tracking · 33 papers
- DVIS++: Improved Decoupled Framework for Universal Video Segmentation (55 citations)
- LVOS: A Benchmark for Large-Scale Long-Term Video Object Segmentation (41 citations)
- EPIC-SOUNDS: A Large-Scale Dataset of Actions That Sound (34 citations)
- JointFormer: A Unified Framework With Joint Modeling for Video Object Segmentation (25 citations)
- EventAid: Benchmarking Event-Aided Image/Video Enhancement Algorithms With Real-Captured Hybrid Dataset (21 citations)
- MeViS: A Multi-Modal Dataset for Referring Motion Expression Video Segmentation (20 citations)
- SMC++: Masked Learning of Unsupervised Video Semantic Compression (19 citations)
- StabStitch++: Unsupervised Online Video Stitching With Spatiotemporal Bidirectional Warps (19 citations)
- BiVM: Accurate Binarized Neural Network for Efficient Video Matting (16 citations)
- Data-Driven Feature Tracking for Event Cameras With and Without Frames (16 citations)
- EvLight++: Low-Light Video Enhancement With an Event Camera: A Large-Scale Real-World Dataset, Novel Method, and More (11 citations)
- Learning High-Quality Dynamic Memory for Video Object Segmentation (9 citations)
and 21 more in the interactive map.
Video-language understanding · 16 papers
- MotionLLM: Understanding Human Behaviors from Human Motions and Videos (116 citations)
- MovieChat+: Question-Aware Sparse Memory for Long Video Question Answering (80 citations)
- Compositional Physical Reasoning of Objects and Events From Videos (65 citations)
- A Survey on Video Temporal Grounding With Multimodal Large Language Model (26 citations)
- Unified Static and Dynamic Network: Efficient Temporal Filtering for Video Grounding (22 citations)
- UniAV: Unified Audio-Visual Perception for Multi-Task Video Event Localization (17 citations)
- VimTS: A Unified Video and Image Text Spotter for Enhancing the Cross-Domain Generalization (6 citations)
- MECD+: Unlocking Event-Level Causal Graph Discovery for Video Reasoning (5 citations)
- PRVR: Partially Relevant Video Retrieval (5 citations)
- Hier-EgoPack: Hierarchical Egocentric Video Understanding With Diverse Task Perspectives (3 citations)
- FiGVCL: Fine-Grained Benchmark and Method for Video Copy Localization (1 citation)
- Video DataFlywheel: Resolving the Impossible Data Trinity in Video-Language Understanding (1 citation)
and 4 more in the interactive map.
Video generation · 15 papers
- MagicTime: Time-Lapse Video Generation Models as Metamorphic Simulators (51 citations)
- Lodge++: High-Quality and Long Dance Generation With Robust Choreography Patterns (34 citations)
- Human Motion Video Generation: A Survey (30 citations)
- MADiff: Motion-Aware Mamba Diffusion Models for Hand Trajectory Prediction on Egocentric Videos (22 citations)
- Interactive Conversational Head Generation (19 citations)
- UniVST: A Unified Framework for Training-Free Localized Video Style Transfer (16 citations)
- Combo: Co-speech holistic 3D human motion generation and efficient customizable adaptation in harmony (10 citations)
- Dubbing Movies via Hierarchical Phoneme Modeling and Acoustic Diffusion Denoising (10 citations)
- Video Diffusion Posterior Sampling for Seeing Beyond Dynamic Scattering Layers (5 citations)
- Creating Multimodal Interactive Digital Twin Characters From Videos: A Dataset and Baseline (3 citations)
- Video4DGen: Enhancing Video and 4D Generation through Mutual Optimization (3 citations)
- Lagrangian Motion Fields for Long-Term Motion Generation (2 citations)
and 3 more in the interactive map.
Action recognition · 10 papers
- Human-Centric Fine-Grained Action Quality Assessment (19 citations)
- Weakly Supervised Micro- and Macro-Expression Spotting Based on Multi-Level Consistency (10 citations)
- CycleACR: Cycle Modeling of Actor-Context Relations for Video Action Detection (9 citations)
- ${\text{CA}^{2}\text{ST}}$: Cross-Attention in Audio, Space, and Time for Holistic Video Recognition (6 citations)
- STADe: Sensory Temporal Action Detection via Temporal-Spectral Representation Learning (5 citations)
- Transferable Unintentional Action Localization With Language-Guided Intention Translation (4 citations)
- Cross-Task Relation-Aware Consistency for Weakly Supervised Temporal Action Detection (2 citations)
- FLAG3D++: A Benchmark for 3D Fitness Activity Comprehension With Language Instruction
- Foundation Model for Skeleton-Based Human Action Understanding
- Heatmap Pooling Network for Action Recognition From RGB Videos
