Research map / Video / Action recognition
Learning a Deep Model for Human Action Recognition from Novel Viewpoints
Hossein Rahmani, Ajmal Saeed Mian, Mubarak Shah
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 |
|---|---|---|---|
| use R-NKTM architecture with virtual-view path instead of canonical-view NKTM | NKTM · IXMAS mean | accuracy (%): 72.5 -> 74.1 | Improved |
| use R-NKTM instead of NKTM | NKTM · UWA3D Multiview Activity II mean | accuracy (%): 63.5 -> 67.4 | Improved |
| use R-NKTM instead of NKTM | NKTM · N-UCLA | accuracy (%): 75.8 -> 78.1 | Improved |
| use R-NKTM instead of NKTM when camera 4 is source or target | NKTM · IXMAS camera 4 | accuracy (%): 54.6 -> 57.8 | Improved |
| use R-NKTM instead of nCTE | nCTE · IXMAS mean | accuracy (%): 67.4 -> 74.1 | Improved |
| use R-NKTM instead of nCTE on UWA3D with view 4 as test | nCTE · UWA3D view 4 test | accuracy (%): 67.5 average, about 8% higher | Improved |
| use R-NKTM instead of DT | DT · IXMAS camera 4 | accuracy (%): 39.8 -> 57.8 | Improved |
| add HOG, HOF, MBH to trajectory cross-view descriptor (ablation) | R-NKTM · IXMAS mean | accuracy (%): 74.1 -> 75.0 | Improved |
| add HOG, HOF, MBH to trajectory cross-view descriptor (ablation) | R-NKTM · IXMAS top view pairs (0-4, 1-4, 3-4) | accuracy (%): 55.2 -> 51.5; 47.0 -> 43.3; 44.2 -> 43.3 | Hurt |
| add HOG, HOF, MBH to trajectory descriptor (ablation) | R-NKTM · UCF Sports | accuracy (%): 76.7 -> 90.0 | Improved |
| use R-NKTM instead of DT with trajectories only | DT · UCF Sports | accuracy (%): 75.2 -> 76.7 | Improved |
| add more concatenated virtual views to cross-view descriptor (ablation) | R-NKTM · IXMAS and UWA3DII | accuracy | Improved |
Closest papers by meaning
- 3D Action Recognition From Novel Viewpoints (CVPR 2016 · similarity 0.87)
- Knowledge Integration Networks for Action Recognition (AAAI 2020 · similarity 0.87)
- A Recurrent Transformer Network for Novel View Action Synthesis (ECCV 2020 · similarity 0.87)
- RPAN: An End-To-End Recurrent Pose-Attention Network for Action Recognition in Videos (ICCV 2017 · similarity 0.86)
- Procedural Generation of Videos to Train Deep Action Recognition Networks (CVPR 2017 · similarity 0.86)
- DVANet: Disentangling View and Action Features for Multi-View Action Recognition (AAAI 2024 · similarity 0.85)
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.
