Research map / Image generation / Editing and style transfer
Human-Centric Image Cropping with Partition-Aware and Content-Preserving Features
Bo Zhang, Li Niu, Xing Zhao, Liqing Zhang
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 partition-aware feature (ablation) | basic region feature · GAICD | SRCC: 0.744 -> 0.774 | Improved |
| add content-preserving feature (ablation) | basic region feature · GAICD | SRCC: 0.744 -> 0.781 | Improved |
| combine partition-aware and content-preserving features (ablation) | basic region feature · GAICD | SRCC: 0.744 -> 0.795 | Improved |
| add human feature to partition-aware feature (ablation) | partition-aware feature · GAICD | SRCC: 0.769 -> 0.774 | Improved |
| add residual learning to partition-aware feature (ablation) | partition-aware feature · GAICD | SRCC: 0.764 -> 0.774 | Improved |
| use nine partitions instead of one (ablation) | partition-aware feature · GAICD | SRCC: 0.746 -> 0.774 | Improved |
| use nine partitions instead of two (ablation) | partition-aware feature · GAICD | SRCC: 0.756 -> 0.774 | Improved |
| use GCN instead of convolution for heatmap (ablation) | content-preserving feature · GAICD | SRCC: 0.762 -> 0.781 | Improved |
| add predicted important-content heatmap (ablation) | content-preserving feature · GAICD | SRCC: 0.741 -> 0.781 | Improved |
| use pseudo ground-truth heatmap instead of saliency map (ablation) | content-preserving feature · GAICD | SRCC: 0.752 -> 0.781 | Improved |
| add region feature to content-preserving feature (ablation) | content-preserving feature · GAICD | SRCC: 0.643 -> 0.781 | Improved |
| use full model on FCDB and FLMS instead of basic (ablation) | Ours(basic) · FCDB+FLMS | IoU: 0.7263 -> 0.7469 | Improved |
Closest papers by meaning
- Find Beauty in the Rare: Contrastive Composition Feature Clustering for Nontrivial Cropping Box Regression (AAAI 2023 · similarity 0.88)
- Image Cropping With Spatial-Aware Feature and Rank Consistency (CVPR 2023 · similarity 0.87)
- Spatial-Semantic Collaborative Cropping for User Generated Content (AAAI 2024 · similarity 0.86)
- Reliable and Efficient Image Cropping: A Grid Anchor Based Approach (CVPR 2019 · similarity 0.86)
- Learning Subject-Aware Cropping by Outpainting Professional Photos (AAAI 2024 · similarity 0.86)
- Cropper: Vision-Language Model for Image Cropping through In-Context Learning (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.
