Research map / Label-efficient and robust learning / Recognition and long-tail
Find Beauty in the Rare: Contrastive Composition Feature Clustering for Nontrivial Cropping Box Regression
Zhiyu Pan, Yinpeng Chen, Jiale Zhang, Hao Lu, Zhiguo Cao, Weicai Zhong
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 C2C to CBLNet (ablation) | CBLNet · FCDB All | IoU: 0.700 -> 0.718 | Improved |
| Add C2C to CBLNet (ablation) | CBLNet · FCDB Few | IoU: 0.391 -> 0.418 | Improved |
| Add C2C to CBLNet (ablation) | CBLNet · FCDB All | BDE: 0.075 -> 0.069 | Improved |
| Add the boundary feature encoder (ablation) | regression baseline · FCDB | IoU: 0.692 -> 0.700 | No clear effect |
| Increase batch size from 4 to 256 (ablation) | C2C · FCDB Few | IoU: 0.424 -> 0.471 | Improved |
| Increase batch size from 4 to 256 (ablation) | C2C · FCDB All | IoU: 0.734 -> 0.750 | Mixed |
| Use regressor re-training (RRT) | CBLNet baseline · FCDB Few | IoU: 0.391 -> 0.416 | Improved |
| Use regressor re-training (RRT) | CBLNet baseline · FCDB Med. | IoU: 0.659 -> 0.653 | Hurt |
| Use LDS&FDS | CBLNet baseline · FCDB Few | IoU: 0.391 -> 0.408 | Improved |
| Use LDS&FDS | CBLNet baseline · FCDB | training time: 0.720 -> 1.187 | Hurt |
Closest papers by meaning
- Learning Subject-Aware Cropping by Outpainting Professional Photos (AAAI 2024 · similarity 0.88)
- Spatial-Semantic Collaborative Cropping for User Generated Content (AAAI 2024 · similarity 0.86)
- Venus: Benchmarking and Empowering Multimodal Large Language Models for Aesthetic Guidance and Cropping (CVPR 2026 · similarity 0.85)
- Cropper: Vision-Language Model for Image Cropping through In-Context Learning (CVPR 2025 · similarity 0.85)
- Can Machines Understand Composition? Dataset and Benchmark for Photographic Image Composition Embedding and Understanding (CVPR 2025 · similarity 0.84)
- $\mathbb{X}$-Sample Contrastive Loss: Improving Contrastive Learning with Sample Similarity Graphs (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.
