Research map / Video / Action recognition
Order-Preserving Wasserstein Discriminant Analysis
Bing Su, Jiahuan Zhou, Ying Wu
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 OWDA dimensionality reduction instead of LSDA before the SVM classifier (few dimensions) | sequence DRS + SVM · MSR Action3D | accuracy and MAP: margin of about 10% | Improved |
| add OWDA reduction to 30 dimensions before the SVM classifier | original 192-dim features + SVM · MSR Action3D | accuracy and MAP: better accuracy and comparable MAP using only 30 dimensions | Improved |
| add OWDA reduction to 25 dimensions before the SVM classifier | original features + SVM · MSR Daily Activity3D | MAP: comparable MAP using only 25 dimensions | No clear effect |
| apply OWDA dimensionality reduction before the SVM classifier | original features + SVM · MSR Daily Activity3D | accuracy | Hurt |
| use OWDA instead of other DRS methods with NN classifier | NN classifier · MSR Daily Activity3D | accuracy | Hurt |
| use OWDA instead of other DRS methods with NN classifier | NN classifier · MSR Daily Activity3D | MAP | Improved |
| use OWDA instead of other DRS methods with SVM classifier | SVM classifier · ChaLearn Gesture | accuracy: margin of about 5% | Improved |
| add OWDA to the original features with NN classifier | original features + NN · ChaLearn Gesture | MAP: margin of 20% | Improved |
| use OWDA instead of LSDA with NN classifier | NN classifier · ChaLearn Gesture | MAP: 5% higher | Improved |
| use OWDA instead of LSDA with NN classifier | NN classifier · ChaLearn Gesture | accuracy | No clear effect |
| use OWDA+SVM instead of LSDA+SVM | ChaLearn pipeline · ChaLearn Gesture (validation) | F-score: 0.767 -> 0.772 | Improved |
| apply OWDA (reduce to 80 dims) before Kernelized-COV | Kernelized-COV · MSR Daily Activity3D | accuracy: 96.3% -> 98.1% | Improved |
Closest papers by meaning
- Linear and Deep Order-Preserving Wasserstein Discriminant Analysis (TPAMI 2021 · similarity 0.96)
- Ratio Trace Formulation of Wasserstein Discriminant Analysis (NeurIPS 2020 · similarity 0.83)
- Conservative Wasserstein Training for Pose Estimation (ICCV 2019 · similarity 0.83)
- Order-Preserving Wasserstein Distance for Sequence Matching (CVPR 2017 · similarity 0.82)
- Tree-Wasserstein Distance for High Dimensional Data with a Latent Feature Hierarchy (ICLR 2025 · similarity 0.82)
- Statistical and Computational Guarantees of Kernel Max-Sliced Wasserstein Distances (ICML 2025 · similarity 0.81)
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.
