Research map / Detection and segmentation / CNN features
Look Closer to See Better: Recurrent Attention Convolutional Neural Network for Fine-Grained Image Recognition
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 |
|---|---|---|---|
| combine features from scales 1, 2, and 3 (ablation) | RA-CNN · CUB-200-2011 | accuracy: 84.7 -> 85.3 | Helped |
| combine features from scales 1, 2, and 3 (ablation) | RA-CNN · Stanford Dogs | accuracy: 86.7 -> 87.3 | Helped |
| combine features from scales 1, 2, and 3 (ablation) | RA-CNN · Stanford Cars | accuracy: 91.8 -> 92.5 | Helped |
| optimize region attention with the pairwise ranking loss (ablation) | RA-CNN · CUB-200-2011 | accuracy: 79.0 -> 82.4 | Helped |
| look at an attended region at scale 2 rather than only the full image (ablation) | RA-CNN · CUB-200-2011 | accuracy: 77.8 -> 82.4 | Helped |
| look at an attended region at scale 2 rather than only the full image (ablation) | RA-CNN · Stanford Dogs | accuracy: 76.7 -> 85.9 | Helped |
| look at an attended region at scale 2 rather than only the full image (ablation) | RA-CNN · Stanford Cars | accuracy: 84.9 -> 90.0 | Helped |
| use scale 3 alone instead of scale 2 alone (ablation) | RA-CNN · CUB-200-2011 | accuracy: 82.4 -> 81.2 | Hurt |
| use scale 3 alone instead of scale 2 alone (ablation) | RA-CNN · Stanford Dogs | accuracy: 85.9 -> 85.0 | Hurt |
| use scale 3 alone instead of scale 2 alone (ablation) | RA-CNN · Stanford Cars | accuracy: 90.0 -> 89.2 | Hurt |
| use recurrent attention and feature learning without bounding-box or part annotations | RA-CNN · CUB-200-2011 | accuracy: 82.6 -> 85.3 | Helped |
| use recurrent attention and feature learning | RA-CNN · Stanford Dogs | accuracy: 84.2 -> 87.3 | Helped |
Closest papers by meaning
- Looking for the Devil in the Details: Learning Trilinear Attention Sampling Network for Fine-Grained Image Recognition (CVPR 2019 · similarity 0.90)
- Attention Convolutional Binary Neural Tree for Fine-Grained Visual Categorization (CVPR 2020 · similarity 0.88)
- Dual Cross-Attention Learning for Fine-Grained Visual Categorization and Object Re-Identification (CVPR 2022 · similarity 0.87)
- Dual Attention Networks for Few-Shot Fine-Grained Recognition (AAAI 2022 · similarity 0.87)
- Multi-scale Activation, Refinement, and Aggregation: Exploring Diverse Cues for Fine-Grained Bird Recognition (AAAI 2025 · similarity 0.87)
- Prompt-CAM: Making Vision Transformers Interpretable for Fine-Grained Analysis (CVPR 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.