Research map / Learning theory and algorithms / Kernels and regression theory
Approximate Log-Hilbert-Schmidt Distances Between Covariance Operators for Image Classification
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 |
|---|---|---|---|
| approximate the Log-HS distance using random Fourier features | covariance-operator image classification · Fish Exp1 | classification accuracy: 53.91%(±4.34) vs 56.74%(±2.87) | No clear effect |
| use quasi-random rather than random Fourier features | Approx LogHS · Fish Exp1 | classification accuracy: 53.91%(±4.34) -> 54.30%(±3.44) | No clear effect |
| use Approx LogHS instead of LogE | image classification · Fish Exp2 | classification accuracy: 46.20%(±1.9) -> 56.2%(±2.2) | Helped |
| use QApprox LogHS instead of LogE | image classification · Fish Exp2 | classification accuracy: 46.20%(±1.9) -> 57.70%(±1.8) | Helped |
| use approximate Log-HS instead of the best reported prior method | image classification · Virus | accuracy: 82.5%(±2.9) [12] vs 81.5%(±2.1) | Hurt |
| use approximate Log-HS instead of the best reported prior method | image classification · KTH-TIPS2b | accuracy: 80.1%(±4.6) [14] -> 83.6%(±5.4) | Helped |
| use approximate Log-HS instead of the best reported prior method | image classification · ETHZ Re-ID Seq1 | accuracy: 90.2%(±1.0) [14] -> 92.0%(±0.3) | Helped |
| increase approximate feature dimension from 200 to 400 (ablation) | Approx LogHS · TinyGraz03 | accuracy: 57% -> 59% | Helped |
| increase approximate feature dimension from 400 to 1000 (ablation) | Approx LogHS · TinyGraz03 | accuracy: 59% -> 60% | Helped |
| increase approximate feature dimension from 200 to 400 (ablation) | QApprox LogHS · TinyGraz03 | accuracy: 55% -> 58% | Helped |
| increase approximate feature dimension from 400 to 1000 (ablation) | QApprox LogHS · TinyGraz03 | accuracy: 58% -> 59% | Helped |
| increase approximate feature dimension from 200 to 1000 (ablation) | Approx LogHS · TinyGraz03 | training time: 3.8s -> 240.4s | Hurt |
Closest papers by meaning
- Intrinsic Riemannian Cross-covariance for Manifold-valued Random Objects (NeurIPS 2026 · similarity 0.86)
- Task-Induced Riemannian Metrics for Vision Transformer Feature Spaces (NeurIPS 2026 · similarity 0.84)
- Euclidean Embedding of Data Using Local Distances (NeurIPS 2026 · similarity 0.83)
- RMLR: Extending Multinomial Logistic Regression into General Geometries (NeurIPS 2024 · similarity 0.83)
- Random features for Grassmannian kernel approximation with bounded rank-one projections (NeurIPS 2026 · similarity 0.83)
- Understanding Matrix Function Normalizations in Covariance Pooling through the Lens of Riemannian Geometry (ICLR 2025 · similarity 0.83)
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.