NeurIPS 2022 accepted papers on Video
58 of the 2,901 papers accepted at NeurIPS 2022 (2.0%) are on Video, compared with 2.0% at NeurIPS 2021. They fall into 4 topics.
See NeurIPS 2022 in the interactive map
Submitting to NeurIPS? 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 · 23 papers
- VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training (435 citations)
- Masked Autoencoders As Spatiotemporal Learners (241 citations)
- Recurrent Video Restoration Transformer with Guided Deformable Attention (81 citations)
- Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos (49 citations)
- VCT: A Video Compression Transformer (37 citations)
- Rethinking Alignment in Video Super-Resolution Transformers (34 citations)
- MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training (28 citations)
- EPIC-KITCHENS VISOR Benchmark: VIdeo Segmentations and Object Relations (24 citations)
- Delving into Sequential Patches for Deepfake Detection (21 citations)
- SCAMPS: Synthetics for Camera Measurement of Physiological Signals (18 citations)
- AnimeSR: Learning Real-World Super-Resolution Models for Animation Videos (13 citations)
- Video compression dataset and benchmark of learning-based video-quality metrics (13 citations)
and 11 more in the interactive map.
Video-language understanding · 15 papers
- ST-Adapter: Parameter-Efficient Image-to-Video Transfer Learning (76 citations)
- Language Models with Image Descriptors are Strong Few-Shot Video-Language Learners (52 citations)
- Egocentric Video-Language Pretraining (45 citations)
- Expectation-Maximization Contrastive Learning for Compact Video-and-Language Representations (35 citations)
- Long-Form Video-Language Pre-Training with Multimodal Temporal Contrastive Learning (33 citations)
- Embracing Consistency: A One-Stage Approach for Spatio-Temporal Video Grounding (15 citations)
- EgoTaskQA: Understanding Human Tasks in Egocentric Videos (13 citations)
- Grounded Video Situation Recognition (6 citations)
- Robustness Analysis of Video-Language Models Against Visual and Language Perturbations (6 citations)
- Scaling Multimodal Pre-Training via Cross-Modality Gradient Harmonization (6 citations)
- LGDN: Language-Guided Denoising Network for Video-Language Modeling (5 citations)
- Multi-modal Grouping Network for Weakly-Supervised Audio-Visual Video Parsing (4 citations)
and 3 more in the interactive map.
Action recognition · 13 papers
- How Transferable are Video Representations Based on Synthetic Data? (16 citations)
- Learning State-Aware Visual Representations from Audible Interactions (13 citations)
- Learning a Condensed Frame for Memory-Efficient Video Class-Incremental Learning (12 citations)
- PointTAD: Multi-Label Temporal Action Detection with Learnable Query Points (11 citations)
- Rethinking Resolution in the Context of Efficient Video Recognition (10 citations)
- Alignment-guided Temporal Attention for Video Action Recognition (9 citations)
- Don't Pour Cereal into Coffee: Differentiable Temporal Logic for Temporal Action Segmentation (6 citations)
- StrokeRehab: A Benchmark Dataset for Sub-second Action Identification (5 citations)
- Look More but Care Less in Video Recognition (4 citations)
- Multi-dataset Training of Transformers for Robust Action Recognition (2 citations)
- Are all Frames Equal? Active Sparse Labeling for Video Action Detection (1 citation)
- Enabling Detailed Action Recognition Evaluation Through Video Dataset Augmentation (1 citation)
and 1 more in the interactive map.
Video generation · 7 papers
- Video Diffusion Models (204 citations)
- Flexible Diffusion Modeling of Long Videos (102 citations)
- MCVD - Masked Conditional Video Diffusion for Prediction, Generation, and Interpolation (45 citations)
- Generating Long Videos of Dynamic Scenes (24 citations)
- Audio-Driven Co-Speech Gesture Video Generation (14 citations)
- NeMF: Neural Motion Fields for Kinematic Animation (12 citations)
- You Never Stop Dancing: Non-freezing Dance Generation via Bank-constrained Manifold Projection (6 citations)
