Research map / 3D vision / Human pose estimation
Rolling Rotations for Recognizing Human Actions From 3D Skeletal Data
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 |
|---|---|---|---|
| Unwrap action curves while rolling along nominal curves instead of using a logarithm map at one point (ablation) | SO3 action-curve representation · Florence3D-Action | accuracy: 86.83 -> 89.82 | Helped |
| Unwrap action curves while rolling along nominal curves instead of using a logarithm map at one point (ablation) | SO3 action-curve representation · MSRAction Pairs | accuracy: 92.96 -> 94.09 | Helped |
| Unwrap action curves while rolling along nominal curves instead of using a logarithm map at one point (ablation) | SO3 action-curve representation · G3D-Gaming | accuracy: 87.82 -> 87.95 | No clear effect |
| Use FTP features instead of concatenating temporal samples (ablation) | Proposed rolling-rotation method · Florence3D-Action | accuracy: 89.82 -> 91.40 | Helped |
| Use FTP features instead of concatenating temporal samples (ablation) | Proposed rolling-rotation method · MSRAction Pairs | accuracy: 94.09 -> 94.67 | Helped |
| Use FTP features instead of concatenating temporal samples (ablation) | Proposed rolling-rotation method · G3D-Gaming | accuracy: 87.95 -> 90.94 | Helped |
| Represent relative body-part geometry with rotations only rather than full rigid transformations | Skeleton action representation · Florence3D-Action | accuracy: 90.71 -> 89.82 | Hurt |
| Represent relative body-part geometry with rotations only rather than full rigid transformations | Skeleton action representation · MSRAction Pairs | accuracy: 93.65 -> 94.09 | Helped |
| Represent relative body-part geometry with rotations only rather than full rigid transformations | Skeleton action representation · G3D-Gaming | accuracy: 91.09 -> 87.95 | Hurt |
Closest papers by meaning
- Deep Learning on Lie Groups for Skeleton-Based Action Recognition (CVPR 2017 · similarity 0.89)
- Revisiting Skeleton-Based Action Recognition (CVPR 2022 · similarity 0.87)
- Skeleton-based Action Recognition with Non-linear Dependency Modeling and Hilbert-Schmidt Independence Criterion (AAAI 2025 · similarity 0.86)
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.85)
- SkeletonGait: Gait Recognition Using Skeleton Maps (AAAI 2024 · similarity 0.84)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (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.