Research map / Video / Action recognition
Tensor Representations via Kernel Linearization for Action Recognition from 3D Skeletons
Piotr Koniusz, Anoop Cherian, Fatih Porikli
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 slice-wise Eigenvalue Power Normalization to SCK (ablation) | SCK · Florence3D-Action | accuracy (%): 88.7 -> 92.98 | Improved |
| add HOSVD Eigenvalue Power Normalization (gamma=0.85) to DCK (ablation) | DCK · Florence3D-Action | accuracy (%): 90.49 -> 92.77 | Improved |
| combine SCK with DCK | SCK · Florence3D-Action | accuracy (%): 92.98 -> 95.23 | Improved |
| combine SCK with DCK | SCK · UTKinect-Action | accuracy (%): 96.08 -> 98.2 | Improved |
| combine SCK with DCK | SCK · MSR-Action3D (actionlets protocol) | accuracy (%): 90.72 -> 91.45 | Improved |
| combine SCK with DCK | SCK · MSR-Action3D (three-subset protocol) | accuracy (%): 93.52 -> 93.96 | Improved |
| use all body-joints (Configuration-I) instead of six-joint subset in DCK (ablation) | DCK · Florence3D-Action | accuracy (%): 93.03 -> 91.65 | Hurt |
| use only two hand joints (Configuration-A) instead of six joints in DCK (ablation) | DCK · Florence3D-Action | accuracy (%): 93.03 -> 88.32 | Hurt |
| set temporal kernel radius sigma3 to 4 (extreme value) (ablation) | SCK · Florence3D-Action | accuracy (%): 92.98 -> 91 | Hurt |
| increase pivots Z2 from 5 to 20 for body-joint subkernel (ablation) | SCK · Florence3D-Action | accuracy (%): Z2=5 gives 92.98; Z2=20 deteriorates slightly | Hurt |
| use DCK Configuration-D instead of Configuration-E (ablation) | DCK · Florence3D-Action | accuracy (%): 92 -> 92.27 | Improved |
Closest papers by meaning
- Tensor Representations for Action Recognition (TPAMI 2021 · similarity 0.93)
- Spatio-Temporal Difference Descriptor for Skeleton-Based Action Recognition (AAAI 2021 · similarity 0.87)
- Subspace Kernel Learning on Tensor Sequences (ICLR 2026 · similarity 0.87)
- Ske2Grid: Skeleton-to-Grid Representation Learning for Action Recognition (ICML 2023 · 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 0.86)
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.
