Research map / Label-efficient and robust learning / Self-supervised and contrastive
Contrastive Learning from Extremely Augmented Skeleton Sequences for Self-Supervised Action Recognition
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace SkeletonCLR with AimCLR | 3-stream skeleton representation learning · NTU-60 xsub | accuracy (%): 75.0 -> 78.9 | Helped |
| Replace SkeletonCLR with AimCLR | 3-stream skeleton representation learning · NTU-60 xview | accuracy (%): 79.8 -> 83.8 | Helped |
| Replace SkeletonCLR with AimCLR | 3-stream skeleton representation learning · PKU-MMD part I | accuracy (%): 85.3 -> 87.8 | Helped |
| Replace SkeletonCLR with AimCLR | 3-stream skeleton representation learning · NTU-120 xsub | accuracy (%): 60.7 -> 68.2 | Helped |
| Replace SkeletonCLR with AimCLR | 3-stream skeleton representation learning · NTU-120 xset | accuracy (%): 62.6 -> 68.8 | Helped |
| Replace normal augmentations with extreme augmentations (ablation) | AimCLR · NTU-60 xsub | accuracy (%): 75.0 -> 71.3 | Hurt |
| Replace normal augmentations with extreme augmentations (ablation) | AimCLR · NTU-60 xview | accuracy (%): 79.8 -> 77.8 | Hurt |
| Add extreme augmentations alongside normal augmentations (ablation) | AimCLR · NTU-60 xsub | accuracy (%): 75.0 -> 77.4 | Helped |
| Add extreme augmentations alongside normal augmentations (ablation) | AimCLR · NTU-60 xview | accuracy (%): 79.8 -> 82.5 | Helped |
| Add EADM to normal and extreme augmentations (ablation) | AimCLR · NTU-60 xsub | accuracy (%): 77.4 -> 78.2 | Helped |
| Add EADM to normal and extreme augmentations (ablation) | AimCLR · NTU-60 xview | accuracy (%): 82.5 -> 82.8 | Helped |
| Add nearest neighbors mining (ablation) | AimCLR with EADM · NTU-60 xsub | accuracy (%): 78.2 -> 78.9 | Helped |
Closest papers by meaning
- MaskCLR: Attention-Guided Contrastive Learning for Robust Action Representation Learning (CVPR 2024 · similarity 0.86)
- Skeleton-based Action Recognition with Non-linear Dependency Modeling and Hilbert-Schmidt Independence Criterion (AAAI 2025 · similarity 0.84)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.84)
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.84)
- SCD-Net: Spatiotemporal Clues Disentanglement Network for Self-Supervised Skeleton-Based Action Recognition (AAAI 2024 · similarity 0.84)
- Stitch, Contrast, and Segment: Learning a Human Action Segmentation Model Using Trimmed Skeleton Videos (AAAI 2025 · similarity 0.83)
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.