Research map / Architectures and efficiency / Attention and transformers
Global Context-Aware Attention LSTM Networks for 3D Action Recognition
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 global-context-aware attention (ablation) | GCA-LSTM · NTU RGB+D X-subject | accuracy: 70.7% -> 74.4% | Helped |
| Add global-context-aware attention (ablation) | GCA-LSTM · NTU RGB+D X-view | accuracy: 79.4% -> 82.8% | Helped |
| Add global-context-aware attention (ablation) | GCA-LSTM · UT-Kinect | accuracy: 97.5% -> 98.5% | Helped |
| Add global-context-aware attention (ablation) | GCA-LSTM · SBU-Kinect Interaction | accuracy | Helped |
| Use 2 attention iterations instead of 1 (ablation) | GCA-LSTM · NTU RGB+D X-subject | accuracy: 71.9% -> 74.4% | Helped |
| Use 2 attention iterations instead of 1 (ablation) | GCA-LSTM · NTU RGB+D X-view | accuracy: 81.1% -> 82.8% | Helped |
| Use 3 attention iterations instead of 2 (ablation) | GCA-LSTM · NTU RGB+D X-subject | accuracy: 74.4% -> 72.7% | Hurt |
| Use 3 attention iterations instead of 2 (ablation) | GCA-LSTM · NTU RGB+D X-view | accuracy: 82.8% -> 81.2% | Hurt |
| Replace the LSTM informativeness gate with soft attention (ablation) | GCA-LSTM · NTU RGB+D | accuracy | Hurt |
Closest papers by meaning
- An Attention Enhanced Graph Convolutional LSTM Network for Skeleton-Based Action Recognition (CVPR 2019 · similarity 0.89)
- Attention in Convolutional LSTM for Gesture Recognition (NeurIPS 2018 · similarity 0.89)
- An End-to-End Spatio-Temporal Attention Model for Human Action Recognition from Skeleton Data (AAAI 2017 · similarity 0.89)
- Regularizing Long Short Term Memory With 3D Human-Skeleton Sequences for Action Recognition (CVPR 2016 · similarity 0.88)
- Co-Occurrence Feature Learning for Skeleton Based Action Recognition Using Regularized Deep LSTM Networks (AAAI 2016 · similarity 0.88)
- Are Spatial-Temporal Graph Convolution Networks for Human Action Recognition Over-Parameterized? (CVPR 2025 · similarity 0.84)
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.