Research map / Label-efficient and robust learning / Recognition and long-tail
HPatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors
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 |
|---|---|---|---|
| Increase the measurement-region scale from ρ=1 to ρ=20 | DoG + SIFT · Oxford Leuven, 1|2 | mAP: 0.31 -> 0.87 | Helped |
| Apply ZCA projection and normalization (ablation) | SIFT · HPatches image matching and patch retrieval | mAP | Helped |
| Apply post-processing normalization (ablation) | DeepCompare descriptors · HPatches patch verification | mAP | Helped |
| Apply post-processing normalization (ablation) | TFeat descriptors · HPatches patch verification | mAP | Helped |
| Use negative patch pairs from the same sequence instead of different sequences | HPatches patch verification · SameSeq versus DiffSeq | mAP | Hurt |
| Evaluate on illumination-change sequences rather than viewpoint-change sequences | HPatches image matching · Illum versus Viewpt | mAP | Hurt |
| Use patch verification alone rather than matching or retrieval | Local descriptor evaluation · HPatches tasks | mAP | Mixed |
| Increase geometric distortion from Easy to Tough | Local descriptors · HPatches tasks | mAP: up to 30% | Hurt |
Closest papers by meaning
- Comparative Evaluation of Hand-Crafted and Learned Local Features (CVPR 2017 · similarity 0.90)
- Local Descriptors Optimized for Average Precision (CVPR 2018 · similarity 0.89)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · similarity 0.85)
- XFeat: Accelerated Features for Lightweight Image Matching (CVPR 2024 · similarity 0.84)
- Optimal Transport Aggregation for Visual Place Recognition (CVPR 2024 · similarity 0.84)
- ETO:Efficient Transformer-based Local Feature Matching by Organizing Multiple Homography Hypotheses (NeurIPS 2024 · 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.