Research map / Video / Video segmentation and tracking
Self-Supervised Video Representation Learning in a Heuristic Decoupled Perspective
Zeen Song, Wenwen Qiang, Changwen Zheng, Hui Xiong, Gang Hua
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add BOD-VCL decoupled static/dynamic contrastive objective | V-SwAV · Static UCF-101 | top-1 accuracy: 64.84 -> 67.57 | Improved |
| add BOD-VCL | V-MoCo · Static UCF-101 | top-1 accuracy: 73.72 -> 76.67 | Improved |
| add BOD-VCL | V-SimCLR · Dynamic UCF-101 | top-1 accuracy: 21.54 -> 24.72 | Improved |
| add BOD-VCL | V-BYOL · Dynamic UCF-101 | top-5 accuracy: 67.94 -> 69.73 | Improved |
| add BOD-VCL | V-BYOL · UCF101 (finetune, R3D-50, 16 frames) | top-1 accuracy: 95.5 -> 96.5 | Improved |
| add BOD-VCL | V-SwAV · UCF101 (finetune, R3D-50) | top-1 accuracy: 87.3 -> 92.0 | Improved |
| add BOD-VCL | V-BYOL · UCF101 (finetune, S3D-G) | top-1 accuracy: 87.8 -> 91.3 | Improved |
| add BOD-VCL | V-BYOL · HMDB51 (finetune, R(2+1)D-18) | top-1 accuracy: 69.4 -> 72.1 | Improved |
| add BOD-VCL | V-BYOL · UCF101 (finetune, ViT-B) | top-1 accuracy: 95.2 -> 96.7 | Improved |
| add BOD-VCL | V-BYOL · SSv2 (ViT-B) | top-1 accuracy: 58.7 -> 63.9 | Improved |
| add BOD-VCL | V-BYOL · Gym99 (R(2+1)D-18) | top-1 accuracy: 88.6 -> 92.7 | Improved |
| add BOD-VCL | V-SimCLR · Kinetics-400 (linear eval) | top-1 accuracy: 60.5 -> 63.8 | Improved |
Closest papers by meaning
- Contrast and Order Representations for Video Self-Supervised Learning (ICCV 2021 · similarity 0.91)
- Motion-Focused Contrastive Learning of Video Representations (ICCV 2021 · similarity 0.89)
- Time-Equivariant Contrastive Video Representation Learning (ICCV 2021 · similarity 0.89)
- Contrastive Learning of Image Representations With Cross-Video Cycle-Consistency (ICCV 2021 · similarity 0.89)
- Probabilistic Representations for Video Contrastive Learning (CVPR 2022 · similarity 0.88)
- Frame-Wise Action Representations for Long Videos via Sequence Contrastive Learning (CVPR 2022 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
