ECCV 2022 accepted papers on Video
199 of the 1,645 papers accepted at ECCV 2022 (12.1%) are on Video, compared with 11.2% at ECCV 2020. They fall into 4 topics.
See ECCV 2022 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 · 96 papers
- XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Model (518 citations)
- Real-Time Intermediate Flow Estimation for Video Frame Interpolation (451 citations)
- FAST-VQA: Efficient End-to-End Video Quality Assessment with Fragment Sampling (299 citations)
- Event-Based Fusion for Motion Deblurring with Cross-Modal Attention (208 citations)
- FILM: Frame Interpolation for Large Motion (199 citations)
- Particle Video Revisited: Tracking through Occlusions Using Point Trajectories (179 citations)
- SeqFormer: Sequential Transformer for Video Instance Segmentation (168 citations)
- Self-Supervised Sparse Representation for Video Anomaly Detection (167 citations)
- Contrast-Phys: Unsupervised Video-Based Remote Physiological Measurement via Spatiotemporal Contrast (131 citations)
- In Defense of Online Models for Video Instance Segmentation (131 citations)
- Video Anomaly Detection by Solving Decoupled Spatio-Temporal Jigsaw Puzzles (130 citations)
- E-NeRV: Expedite Neural Video Representation with Disentangled Spatial-Temporal Context (121 citations)
and 84 more in the interactive map.
Action recognition · 49 papers
- ActionFormer: Localizing Moments of Actions with Transformers (525 citations)
- Expanding Language-Image Pretrained Models for General Video Recognition (410 citations)
- TALLFormer: Temporal Action Localization with a Long-Memory Transformer (143 citations)
- Long Movie Clip Classification with State-Space Video Models (109 citations)
- Unified Fully and Timestamp Supervised Temporal Action Segmentation via Sequence to Sequence Translation (108 citations)
- ReAct: Temporal Action Detection with Relational Queries (97 citations)
- Real-Time Online Video Detection with Temporal Smoothing Transformers (96 citations)
- Temporal Lift Pooling for Continuous Sign Language Recognition (91 citations)
- Dual-Evidential Learning for Weakly-Supervised Temporal Action Localization (82 citations)
- Action Quality Assessment with Temporal Parsing Transformer (77 citations)
- Few-Shot Action Recognition with Hierarchical Matching and Contrastive Learning (73 citations)
- Compound Prototype Matching for Few-Shot Action Recognition (66 citations)
and 37 more in the interactive map.
Video-language understanding · 31 papers
- Prompting Visual-Language Models for Efficient Video Understanding (458 citations)
- Frozen CLIP Models Are Efficient Video Learners (242 citations)
- TS2-Net: Token Shift and Selection Transformer for Text-Video Retrieval (166 citations)
- Video Graph Transformer for Video Question Answering (110 citations)
- Learning Audio-Video Modalities from Image Captions (82 citations)
- "Fine-Grained Egocentric Hand-Object Segmentation: Dataset, Model, and Applications" (70 citations)
- LocVTP: Video-Text Pre-training for Temporal Localization (54 citations)
- Lightweight Attentional Feature Fusion: A New Baseline for Text-to-Video Retrieval (51 citations)
- ECLIPSE: Efficient Long-Range Video Retrieval Using Sight and Sound (46 citations)
- COMPOSER: Compositional Reasoning of Group Activity in Videos with Keypoint-Only Modality (42 citations)
- MILES: Visual BERT Pre-training with Injected Language Semantics for Video-Text Retrieval (39 citations)
- The Anatomy of Video Editing: A Dataset and Benchmark Suite for AI-Assisted Video Editing (39 citations)
and 19 more in the interactive map.
Video generation · 23 papers
- TEMOS: Generating Diverse Human Motions from Textual Descriptions (438 citations)
- MotionCLIP: Exposing Human Motion Generation to CLIP Space (400 citations)
- TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Texts (263 citations)
- Text2LIVE: Text-Driven Layered Image and Video Editing (252 citations)
- NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion (221 citations)
- StyleHEAT: One-Shot High-Resolution Editable Talking Face Generation via Pre-trained StyleGAN (202 citations)
- BEAT: A Large-Scale Semantic and Emotional Multi-modal Dataset for Conversational Gestures Synthesis (191 citations)
- CelebV-HQ: A Large-Scale Video Facial Attributes Dataset (189 citations)
- Long Video Generation with Time-Agnostic VQGAN and Time-Sensitive Transformer (180 citations)
- PoseGPT: Quantization-Based 3D Human Motion Generation and Forecasting (80 citations)
- Responsive Listening Head Generation: A Benchmark Dataset and Baseline (70 citations)
- Implicit Neural Representations for Variable Length Human Motion Generation (67 citations)
and 11 more in the interactive map.
