Research map / Video / Action recognition
Geometric Deep Neural Network Using Rigid and Non-Rigid Transformations for Human Action Recognition
Rasha Friji, Hassen Drira, Faten Chaieb, Hamza Kchok, Sebastian Kurtek
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 transformation layer to baseline CNN-LSTM (ablation) | CNN-LSTM baseline · NTU-RGB+D (CS) | accuracy: 85.1 -> 89.6 | Improved |
| add transformation layer to baseline CNN-LSTM (ablation) | CNN-LSTM baseline · NTU-RGB+D (CV) | accuracy: 91.2 -> 91.5 | No clear effect |
| add transformation layer to baseline CNN-LSTM (ablation) | CNN-LSTM baseline · NTU-RGB+D120 | accuracy: 56.0 -> 57.2 (CS); 63.5 -> 63.8 (Cset) | No clear effect |
| add tangent space projection to baseline (ablation) | CNN-LSTM baseline · NTU-RGB+D (CS) | accuracy: 85.1 -> 94.1 | Improved |
| add tangent space projection to baseline (ablation) | CNN-LSTM baseline · NTU-RGB+D120 (CS) | accuracy: 56.0 -> 63.9 | Improved |
| add transformation layer on top of tangent projection (ablation) | KShapeNet · NTU-RGB+D120 (CS) | accuracy: 63.9 -> 90.6 | Improved |
| add transformation layer on top of tangent projection (ablation) | KShapeNet · NTU-RGB+D (CS) | accuracy: 94.1 -> 97.0 | Improved |
| use cubic-spline interpolation instead of zero padding (ablation) | KShapeNet · NTU-RGB+D (CS) | accuracy: 81.3% -> 97.0% | Improved |
| use cubic-spline interpolation instead of zero padding (ablation) | KShapeNet · NTU-RGB+D (CV) | accuracy: 85.1% -> 98.5% | Improved |
| use angle-based instead of matrix variant for non-rigid transformations (ablation) | KShapeNet · NTU-RGB+D (CV) | accuracy: 96.9 -> 98.5 | Improved |
| use angle-based instead of matrix variant for non-rigid transformations (ablation) | KShapeNet · NTU-RGB+D120 (Cset) | accuracy: 84.3 -> 86.7 | Improved |
| use matrix-based instead of angle-based variant for rigid transformations (ablation) | KShapeNet · NTU-RGB+D (CV) | accuracy: 96.3 -> 97.1 | Improved |
Closest papers by meaning
- A New Representation of Skeleton Sequences for 3D Action Recognition (CVPR 2017 · similarity 0.90)
- Topology-Aware Convolutional Neural Network for Efficient Skeleton-Based Action Recognition (AAAI 2022 · similarity 0.89)
- Graph CNNs with Motif and Variable Temporal Block for Skeleton-Based Action Recognition (AAAI 2019 · similarity 0.88)
- Geometric Deep Neural Network Using Rigid and Non-rigid Transformations for Landmark-based Human Behavior Analysis (TPAMI 2023 · similarity 0.88)
- Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition (AAAI 2024 · similarity 0.88)
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.87)
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.
