ICCV 2021 accepted papers on Video
187 of the 1,612 papers accepted at ICCV 2021 (11.6%) are on Video, compared with 9.7% at ICCV 2019. They fall into 4 topics.
See ICCV 2021 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 segmentation and tracking · 91 papers
- Weakly-Supervised Video Anomaly Detection With Robust Temporal Feature Magnitude Learning (662 citations)
- Learning To Estimate Hidden Motions With Global Motion Aggregation (413 citations)
- A Hybrid Video Anomaly Detection Framework via Memory-Augmented Flow Reconstruction and Flow-Guided Frame Prediction (395 citations)
- Exploring Temporal Coherence for More General Video Face Forgery Detection (352 citations)
- ID-Reveal: Identity-Aware DeepFake Video Detection (238 citations)
- Visual Alignment Constraint for Continuous Sign Language Recognition (232 citations)
- XVFI: eXtreme Video Frame Interpolation (204 citations)
- TF-Blender: Temporal Feature Blender for Video Object Detection (194 citations)
- Spatial-Temporal Transformer for Dynamic Scene Graph Generation (178 citations)
- Self-Mutual Distillation Learning for Continuous Sign Language Recognition (167 citations)
- The Way to My Heart Is Through Contrastive Learning: Remote Photoplethysmography From Unlabelled Video (166 citations)
- Asymmetric Bilateral Motion Estimation for Video Frame Interpolation (161 citations)
and 79 more in the interactive map.
Action recognition · 58 papers
- ViViT: A Video Vision Transformer (3,115 citations)
- TAM: Temporal Adaptive Module for Video Recognition (409 citations)
- Gait Recognition via Effective Global-Local Feature Representation and Local Temporal Aggregation (375 citations)
- Anticipative Video Transformer (240 citations)
- VidTr: Video Transformer Without Convolutions (231 citations)
- Relaxed Transformer Decoders for Direct Action Proposal Generation (213 citations)
- Context-Sensitive Temporal Feature Learning for Gait Recognition (205 citations)
- Evidential Deep Learning for Open Set Action Recognition (192 citations)
- GroupFormer: Group Activity Recognition With Clustered Spatial-Temporal Transformer (169 citations)
- 3D Local Convolutional Neural Networks for Gait Recognition (168 citations)
- Video Self-Stitching Graph Network for Temporal Action Localization (165 citations)
- Group-Aware Contrastive Regression for Action Quality Assessment (163 citations)
and 46 more in the interactive map.
Video-language understanding · 22 papers
- Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval (961 citations)
- Just Ask: Learning To Answer Questions From Millions of Narrated Videos (235 citations)
- End-to-End Dense Video Captioning With Parallel Decoding (206 citations)
- HiT: Hierarchical Transformer With Momentum Contrast for Video-Text Retrieval (175 citations)
- TeachText: CrossModal Generalized Distillation for Text-Video Retrieval (152 citations)
- Fast Video Moment Retrieval (134 citations)
- TACo: Token-Aware Cascade Contrastive Learning for Video-Text Alignment (123 citations)
- Adaptive Focus for Efficient Video Recognition (120 citations)
- Cross-Sentence Temporal and Semantic Relations in Video Activity Localisation (92 citations)
- Motion Guided Region Message Passing for Video Captioning (81 citations)
- STVGBert: A Visual-Linguistic Transformer Based Framework for Spatio-Temporal Video Grounding (69 citations)
- Zero-Shot Natural Language Video Localization (62 citations)
and 10 more in the interactive map.
Video generation · 16 papers
- Action-Conditioned 3D Human Motion Synthesis With Transformer VAE (574 citations)
- Audio2Gestures: Generating Diverse Gestures From Speech Audio With Conditional Variational Autoencoders (150 citations)
- FACIAL: Synthesizing Dynamic Talking Face With Implicit Attribute Learning (146 citations)
- Speech Drives Templates: Co-Speech Gesture Synthesis With Learned Templates (79 citations)
- A New Journey From SDRTV to HDRTV (70 citations)
- Dynamic Surface Function Networks for Clothed Human Bodies (54 citations)
- Neural Video Portrait Relighting in Real-Time via Consistency Modeling (49 citations)
- Internal Video Inpainting by Implicit Long-Range Propagation (42 citations)
- Long-Term Temporally Consistent Unpaired Video Translation From Simulated Surgical 3D Data (42 citations)
- SLAMP: Stochastic Latent Appearance and Motion Prediction (32 citations)
- Learning To Cut by Watching Movies (22 citations)
- A Light Stage on Every Desk (20 citations)
and 4 more in the interactive map.
