Research map / Video / Action recognition
Towards To-a-T Spatio-Temporal Focus 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 gradient-guided spatio-temporal focus and STF modules to MS-G3D | MS-G3D · NTU-60, NTU-120, and Kinetics-400 benchmarks; all reported modalities and splits | top-1 accuracy (%): 2.46%; 2.79%; 2.64% | Helped |
| add STF coherence loss (ablation) | MS-G3D, NTU-60 x-sub J · NTU-60 x-sub, J | top-1 accuracy (%): 88.77 -> 89.33; ↑ 0.56 | Helped |
| add STF divergence loss to the coherence-loss variant (ablation) | MS-G3D, NTU-60 x-sub J · NTU-60 x-sub, J | top-1 accuracy (%): 89.33 -> 89.92; ↑ 1.15 over MS-G3D | Helped |
| add STF adjacency loss to the coherence-and-divergence variant (ablation) | MS-G3D, NTU-60 x-sub J · NTU-60 x-sub, J | top-1 accuracy (%): 89.92 -> 90.65; ↑ 1.88 over MS-G3D | Helped |
| add STF exploration loss to coherence, divergence, and adjacency losses (ablation) | MS-G3D, NTU-60 x-sub J · NTU-60 x-sub, J | top-1 accuracy (%): 90.65 -> 91.34; ↑ 2.57 over MS-G3D | Helped |
| train STF rather than MS-G3D with 10/20/25% of NTU-60 x-sub training data | MS-G3D · NTU-60 x-sub, J, scarce-data training | top-1 accuracy (%): 72.01 -> 72.73; 79.11 -> 80.77; 81.91 -> 84.27 | Helped |
| transfer the pretrained model from NTU-60 and fine-tune on NTU-120 | MS-G3D · NTU-120 x-sub, dataset shift | top-1 accuracy (%): 82.34% vs. 83.38%; 1.04% | Helped |
Closest papers by meaning
- Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition (AAAI 2018 · similarity 0.93)
- Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition (AAAI 2024 · similarity 0.90)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.90)
- Spatio-Temporal Fusion for Human Action Recognition via Joint Trajectory Graph (AAAI 2024 · similarity 0.90)
- BlockGCN: Redefine Topology Awareness for Skeleton-Based Action Recognition (CVPR 2024 · similarity 0.90)
- Are Spatial-Temporal Graph Convolution Networks for Human Action Recognition Over-Parameterized? (CVPR 2025 · similarity 0.90)
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.