Research map / Label-efficient and robust learning / Recognition and long-tail
DuoCLR: Dual-Surrogate Contrastive Learning for Skeleton-based Human Action Segmentation
Haitao Tian
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 CPC and ROR together instead of ROR only (ablation) | DuoCLR · PKU-MMD Part 1 CS (linear) | accuracy: 76.7 -> 85.2 | Improved |
| add ROR to CPC (ablation) | DuoCLR · PKU-MMD Part 1 CV (linear) | accuracy: 85.1 -> 84.7 | Mixed |
| add Warp to Shuffle (ablation) | DuoCLR · PKU-MMD Part 1 CS | accuracy: 73.8 -> 85.2 | Improved |
| replace Shear/Crop with Shuffle and Warp (ablation) | DuoCLR · PKU-MMD Part 1 CS | accuracy: 58.3 -> 85.2 | Improved |
| add Shear and Crop to Shuffle and Warp (ablation) | DuoCLR · PKU-MMD Part 1 CS | accuracy: 85.2 -> 87.1 | Improved |
| use supervised labels rather than unsupervised proxies (ablation) | DuoCLR · PKU-MMD Part 1 CS | accuracy: 81.6 -> 85.2 | Improved |
| fine-tune DuoCLR pretraining vs training from scratch | DuoCLR · PKU-MMD Part 2 CS | accuracy: 69.3 -> 82.6 | Improved |
| linear eval DuoCLR vs supervised pretraining baseline | DuoCLR · PKU-MMD Part 1 CS | mAP@0.1: 56.0 -> 85.2 | Improved |
| pretrain DuoCLR on trimmed PKU1 instead of training Baseline-II (ablation) | DuoCLR · PKU-MMD Part 2 CS | accuracy: 69.3 -> 78.1 | Improved |
| increase labeled target data from 5% to 10% | DuoCLR (fine-tune) · PKU-MMD Part 1 CS | accuracy: 74.6 -> 81.9 | Improved |
| DuoCLR vs LAC in TSU fine-tune | DuoCLR · TSU CS | mAP: 36.8 -> 35.1 | Hurt |
| increase action granularity G from 1 to about 4 (ablation) | DuoCLR · PKU-MMD Part 2 | accuracy | Improved |
Closest papers by meaning
- Hierarchical Consistent Contrastive Learning for Skeleton-Based Action Recognition with Growing Augmentations (AAAI 2023 · similarity 0.90)
- Hierarchical Contrast for Unsupervised Skeleton-Based Action Representation Learning (AAAI 2023 · similarity 0.90)
- Stitch, Contrast, and Segment: Learning a Human Action Segmentation Model Using Trimmed Skeleton Videos (AAAI 2025 · similarity 0.89)
- MaskCLR: Attention-Guided Contrastive Learning for Robust Action Representation Learning (CVPR 2024 · similarity 0.88)
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.86)
- USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelation (AAAI 2025 · similarity 0.86)
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.
