Research map / Video / Action recognition
Foundation Model for Skeleton-Based Human Action Understanding
Hongsong Wang, Wanjiang Weng, Junbo Wang, Fang Zhao, Guo-Sen Xie, Xin Geng, Liang Wang
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 feature decorrelation (VAC + MG-FD) in place of negative-based pretraining (ablation) | USDRL STTR · NTU-60 x-sub | recognition accuracy: 74.8 -> 83.7 | Improved |
| add feature decorrelation (VAC + MG-FD) in place of negative-based pretraining (ablation) | USDRL STTR · NTU-60 x-sub | retrieval accuracy: 63.8 -> 73.2 | Improved |
| add cross-correlation (XC) term (ablation) | USDRL STTR · NTU-60 x-sub | recognition accuracy: 83.7 -> 84.2 | Improved |
| add cross-correlation (XC) term (ablation) | USDRL DSTE · NTU-60 x-sub | recognition accuracy: 84.8 -> 85.2 | Improved |
| add multi-grained modeling (MG-FD) (ablation) | USDRL DSTE · NTU-60 x-sub | recognition accuracy: 84.6 -> 85.2 | Improved |
| replace STTR encoder with DSTE encoder (ablation) | USDRL · NTU-60 x-sub | recognition accuracy: 84.2 -> 85.2 | Improved |
| use DSTE rather than STTR backbone | USDRL · PKU-MMD I action detection | mAP_a: 66.1 -> 75.7 | Improved |
| use only DSA or only CA module (alpha 0 or 1) (ablation) | DSTE · NTU-60 x-sub | accuracy | Hurt |
| increase encoder layers from 1 to 2 (ablation) | DSTE · NTU-60 x-sub | recognition accuracy: 84.5 -> 85.2 | Improved |
| increase encoder layers from 2 to 3 (ablation) | DSTE · NTU-60 x-sub | recognition accuracy: 85.2 -> 84.9 | No clear effect |
| ensemble three modalities (joint, motion, bone) | USDRL STTR · NTU-60 x-sub | recognition accuracy: 84.2 -> 85.8 | Improved |
| use joint-only USDRL instead of 3s-CMD | 3s-CMD · NTU-60 x-sub | recognition accuracy: 84.1 -> 84.2 | Improved |
Closest papers by meaning
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.92)
- USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelation (AAAI 2025 · similarity 0.90)
- Neuron: Learning Context-Aware Evolving Representations for Zero-Shot Skeleton Action Recognition (CVPR 2025 · similarity 0.90)
- Stitch, Contrast, and Segment: Learning a Human Action Segmentation Model Using Trimmed Skeleton Videos (AAAI 2025 · similarity 0.89)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.89)
- Part-aware Unified Representation of Language and Skeleton for Zero-shot Action Recognition (CVPR 2024 · similarity 0.89)
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.
