Research map / Detection and segmentation / CNN features
Fine-Grained Recognition as HSnet Search for Informative Image Parts
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 |
|---|---|---|---|
| Use HSnet sequential search over multiple image-part proposals rather than direct CNN classification | CUB-2011 · CUB-2011 | top-1 accuracy: 82.3 -> 87.5 | Helped |
| Use HSnet sequential search over multiple image-part proposals rather than direct CNN classification | Cars-196 · Cars-196 | top-1 accuracy: 88.5 -> 93.9 | Helped |
| Add sequential reasoning over one bounding box rather than classifying directly from an image (ablation) | CUB-2011 · CUB-2011 | top-1 accuracy: 82.3 -> 85.7 | Helped |
| Add sequential reasoning over one bounding box rather than classifying directly from an image (ablation) | Cars-196 · Cars-196 | top-1 accuracy: 88.5 -> 92.2 | Helped |
| Use one ground-truth box at each step rather than a single direct CNN prediction (ablation) | CUB-2011 · CUB-2011 | top-1 accuracy: 82.3 -> 86.2 | Helped |
| Replace one sequentially searched box with multiple proposals per search state (ablation) | HSnet · CUB-2011 | top-1 accuracy: 85.7 -> 87.5 | Helped |
| Replace one sequentially searched box with multiple proposals per search state (ablation) | HSnet · Cars-196 | top-1 accuracy: 92.2 -> 93.9 | Helped |
| Use ground-truth parts as the one-box sequence rather than one searched box (ablation) | CUB-2011 · CUB-2011 | top-1 accuracy: 85.7 -> 86.2 | Helped |
Closest papers by meaning
- Part-Stacked CNN for Fine-Grained Visual Categorization (CVPR 2016 · similarity 0.90)
- Fine-Grained Image Classification via Combining Vision and Language (CVPR 2017 · similarity 0.89)
- Interpretable and Accurate Fine-grained Recognition via Region Grouping (CVPR 2020 · similarity 0.89)
- Fine-Grained Recognition: Accounting for Subtle Differences between Similar Classes (AAAI 2020 · similarity 0.89)
- Democratizing Fine-grained Visual Recognition with Large Language Models (ICLR 2024 · similarity 0.86)
- Fine-grained Classes and How to Find Them (ICML 2024 · 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.