Research map / Detection and segmentation / CNN features
Mining Discriminative Triplets of Patches for Fine-Grained 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 |
|---|---|---|---|
| add order and shape geometric constraints to appearance-based triplet localization (ablation) | Triplet detector · FG3DCar | triplet localization accuracy (%): 24.9 -> 35.3 | Helped |
| add the order constraint (ablation) | Triplet detector · FG3DCar | triplet localization accuracy (%): 24.9 -> 27.7 | Helped |
| add the shape constraint (ablation) | Triplet detector · FG3DCar | triplet localization accuracy (%): 24.9 -> 34.4 | Helped |
| add geometric constraints (ablation) | BoT with HOG features · BMVC-14 | accuracy (%): 94.1 -> 96.6 | Helped |
| add geometric constraints (ablation) | BoT with HOG features · Cars-196 | accuracy (%): 84.6 -> 85.7 | Helped |
| add geometric constraints (ablation) | BoT with VGGNet-16 features · Cars-196 | accuracy (%): 91.2 -> 92.5 | Helped |
| add geometric constraints (ablation) | BoT with VGGNet-16 features · FGVC-Aircraft | accuracy (%): 86.7 -> 88.4 | Helped |
| use 10 triplets per class instead of the reported 300-triplet setting (ablation) | BoT with geometric constraints · BMVC-14 | accuracy (%): 84.9 | Mixed |
| increase the number of triplets per class beyond 300 (ablation) | BoT with geometric constraints · BMVC-14 | accuracy (%) | Hurt |
Closest papers by meaning
- Interpretable and Accurate Fine-grained Recognition via Region Grouping (CVPR 2020 · similarity 0.85)
- Recognize Any Regions (NeurIPS 2024 · similarity 0.83)
- Twofold Debiasing Enhances Fine-Grained Learning with Coarse Labels (AAAI 2025 · similarity 0.83)
- Novel Class Discovery for Ultra-Fine-Grained Visual Categorization (CVPR 2024 · similarity 0.83)
- MoRe: Class Patch Attention Needs Regularization for Weakly Supervised Semantic Segmentation (AAAI 2025 · similarity 0.83)
- Fine-grained Classes and How to Find Them (ICML 2024 · similarity 0.82)
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.