Research map / Graph learning / Knowledge and temporal graphs
Spatio-Temporal Graph Convolution for Skeleton Based 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 |
|---|---|---|---|
| add one-hop neighborhood filtering (ablation) | STGC · NTU RGB+D Cross View | accuracy: 80.87% -> 81.11% | Helped |
| use multi-scale filtering with independent signals instead of the single-scale STGC configuration (ablation) | STGC · NTU RGB+D Cross View | accuracy: 81.11% -> 79.71% | Hurt |
| model interactions between signals (ablation) | STGC_K · NTU RGB+D Cross View | accuracy: 79.71% -> 81.84% | Helped |
| stack two STGC_K layers (ablation) | STGC_K (indep.) · NTU RGB+D Cross View | accuracy: 79.71% -> 83.57% | Helped |
| stack two STGC_K layers (ablation) | STGC_K (dep.) · NTU RGB+D Cross View | accuracy: 81.84% -> 86.28% | Helped |
| combine spatial and temporal filtering rather than using either alone (ablation) | STGC_K · HDM05 Protocol 1 | accuracy | Helped |
| increase temporal and spatial kernel scales; best reported setting is K1=2 and K2=6 (ablation) | STGC_K · HDM05 Protocol 1 | accuracy | Mixed |
| use Deep STGC_K instead of STGC_K (ablation) | STGC_K · Florence 3D | accuracy: 97.67% -> 99.07% | Helped |
| use Deep STGC_K instead of STGC_K (ablation) | STGC_K · HDM05 Protocol 2 | accuracy: 83.40%$\pm$1.30 -> 85.29%$\pm$1.33 | Helped |
| use Deep STGC_K instead of STGC_K (ablation) | STGC_K · LSC Cross Subject | precision: 84.0% -> 85.4% | Helped |
| use Deep STGC_K instead of STGC_K (ablation) | STGC_K · NTU RGB+D Cross Subject | accuracy: 72.09% -> 74.85% | Helped |
| use STGC_K (dep.) instead of STGC_K (indep.) (ablation) | STGC_K · Florence 3D | accuracy: 98.14% -> 97.67% | Hurt |
Closest papers by meaning
- Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition (AAAI 2018 · similarity 0.95)
- Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition (AAAI 2024 · similarity 0.92)
- Spatio-Temporal Fusion for Human Action Recognition via Joint Trajectory Graph (AAAI 2024 · similarity 0.90)
- Are Spatial-Temporal Graph Convolution Networks for Human Action Recognition Over-Parameterized? (CVPR 2025 · similarity 0.90)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.89)
- BlockGCN: Redefine Topology Awareness for Skeleton-Based 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.