Research map / Graph learning / Knowledge and temporal graphs
Skeleton-Based Action Recognition With Shift Graph Convolutional Network
Check 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 non-local spatial shift to the point-wise baseline (ablation) | Spatial point-wise baseline · NTU RGB+D X-view | Top-1 accuracy: 90.9 -> 94.5 | Helped |
| Use non-local rather than local spatial shift (ablation) | Spatial shift GCN · NTU RGB+D X-view | Top-1 accuracy: 93.9 -> 94.5 | Helped |
| Add a learnable mask to the non-local spatial shift (ablation) | Non-local shift graph convolution · NTU RGB+D X-view | Top-1 accuracy: 94.2 -> 94.5 | Helped |
| Replace multiple adjacency matrices with one (ablation) | Regular spatial GCN · NTU RGB+D X-view | Top-1 accuracy: 93.4 -> 92.1 | Hurt |
| Replace multiple adjacency matrices with one (ablation) | Adaptive GCN · NTU RGB+D X-view | Top-1 accuracy: 93.9 -> 92.9 | Hurt |
| Add spatial shift to a spatial point-wise convolution (ablation) | Spatial point-wise baseline · NTU RGB+D X-view | Top-1 accuracy: 90.9 -> 93.5 | Helped |
| Use Shift+Conv+Shift instead of Shift+Conv (ablation) | Spatial shift graph convolution · NTU RGB+D X-view | Top-1 accuracy: 93.5 -> 93.9; 94.0 -> 94.2 | Helped |
| Add temporal shift to temporal point-wise convolution (ablation) | Temporal point-wise baseline · NTU RGB+D X-view | Top-1 accuracy: 79.2 -> 94.2 | Helped |
| Use adaptive rather than manually set temporal receptive fields (ablation) | Temporal shift convolution · NTU RGB+D X-view | Top-1 accuracy: 93.7 -> 94.2 | Helped |
| Use Shift+Conv+Shift rather than Shift+Conv (ablation) | Adaptive temporal shift convolution · NTU RGB+D X-view | Top-1 accuracy: 94.0 -> 94.2 | Helped |
| Replace regular spatial convolution with spatial shift (ablation) | ST-GCN · NTU RGB+D X-view | Top-1 accuracy: 93.4 -> 94.5 | Helped |
| Replace regular temporal convolution with temporal shift (ablation) | ST-GCN · NTU RGB+D X-view | Top-1 accuracy: 93.4 -> 94.2 | Helped |
Closest papers by meaning
- Multi-Scale Spatial Temporal Graph Convolutional Network for Skeleton-Based Action Recognition (AAAI 2021 · similarity 0.94)
- Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition (AAAI 2024 · similarity 0.94)
- BlockGCN: Redefine Topology Awareness for Skeleton-Based Action Recognition (CVPR 2024 · similarity 0.91)
- Are Spatial-Temporal Graph Convolution Networks for Human Action Recognition Over-Parameterized? (CVPR 2025 · similarity 0.91)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.91)
- Spatio-Temporal Fusion for Human Action Recognition via Joint Trajectory Graph (AAAI 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.