Research map / Learning theory and algorithms / Matrix and tensor methods
Learning Discriminative ab-Divergences for Positive Definite Matrices
Anoop Cherian, Panagiotis Stanitsas, Mehrtash Harandi, Vassilios Morellas, Nikolaos Papanikolopoulos
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 |
|---|---|---|---|
| learn the divergence and dictionary jointly (IDDL) instead of LEML | SPD classifier · JHMDB | accuracy: 58.85% -> 68.3% | Improved |
| learn the divergence and dictionary jointly (IDDL) instead of kernel sparse coding with log-Euclidean | SPD classifier · 3D Object (SHREC) | accuracy: 59.26% -> 96.08% | Improved |
| learn the divergence and dictionary jointly (IDDL) instead of LEML | SPD classifier · BRODATZ | accuracy: 47.15% -> 74.10% | Improved |
| use IDDL instead of linear SVM on log-Euclidean mapped matrices | SPD classifier · BRODATZ | accuracy: 86.80% -> 74.10% | Hurt |
| use IDDL instead of linear SVM on log-Euclidean mapped matrices | SPD classifier · 3D Object (SHREC) | accuracy: 98.9% -> 96.08% | Hurt |
| use IDDL instead of linear SVM on log-Euclidean mapped matrices | SPD classifier · VIRUS | accuracy: 68.00% -> 78.39% | Improved |
| use IDDL instead of 1-NN with JBLD | SPD classifier · Myometrium Cancer | accuracy: 93.20% -> 94.66% | Improved |
| learn alpha and beta instead of fixing them (ablation) | IDDL · KTH-TIPS2 and VIRUS | accuracy | Improved |
| learn separate alpha, beta per atom (IDDL-N, IDDL-V) instead of one shared pair (IDDL-S) (ablation) | IDDL · six datasets | accuracy | Improved |
| increase the dictionary size beyond 5C atoms for C classes (ablation) | IDDL · eight datasets | accuracy | No clear effect |
Closest papers by meaning
- Learning Log-Determinant Divergences for Positive Definite Matrices (TPAMI 2021 · similarity 0.95)
- Revisiting Metric Learning for SPD Matrix Based Visual Representation (CVPR 2017 · similarity 0.88)
- Dimensionality Reduction on SPD Manifolds: The Emergence of Geometry-Aware Methods (TPAMI 2017 · similarity 0.87)
- Connecting Jensen–Shannon and Kullback–Leibler Divergences: A New Bound for Representation Learning (NeurIPS 2025 · similarity 0.85)
- Task-Induced Riemannian Metrics for Vision Transformer Feature Spaces (NeurIPS 2026 · similarity 0.85)
- Discriminating image representations with principal distortions (ICLR 2025 · 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.
