Research map / Detection and segmentation / Open-set and few-shot detection
F-BRS: Rethinking Backpropagating Refinement for Interactive Segmentation
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 f-BRS-B rather than the strongest listed prior baseline | interactive segmentation · GrabCut | NoC@90: 3.58 -> 2.46 | Helped |
| use f-BRS-B rather than the strongest listed prior baseline | interactive segmentation · Berkeley | NoC@90: 5.60 -> 4.34 | Helped |
| use f-BRS-B rather than the strongest listed prior baseline | interactive segmentation · SBD | NoC@90: 10.78 -> 7.73 | Helped |
| use f-BRS-B rather than the strongest listed prior baseline | interactive segmentation · DAVIS | NoC@90: 9.57 -> 7.41 | Helped |
| replace input-image optimization with channel-wise scale and bias optimization at the chosen feature layer | ResNet-50 interactive segmentation · Berkeley | seconds per click: 1.117 -> 0.132 | Helped |
| replace input-image optimization with channel-wise scale and bias optimization at the chosen feature layer | ResNet-50 interactive segmentation · DAVIS | NoC@90: 7.58 -> 7.81 | Mixed |
| add Zoom-In (ablation) | f-BRS-B · Berkeley | NoC@90: 5.69 -> 4.34 | Helped |
| add Zoom-In (ablation) | f-BRS-B · GrabCut | NoC@90: 2.68 -> 2.98 | Hurt |
| add feature-scale and bias optimization rather than using feed-forward inference alone (ablation) | interactive segmentation · Berkeley | NoC@90: 5.18 -> 4.34 | Helped |
| add feature-scale and bias optimization rather than using feed-forward inference alone (ablation) | interactive segmentation · DAVIS | NoC@90: 8.18 -> 7.81 | Helped |
| move the optimized scale and bias to after the backbone instead of the f-BRS-B location (ablation) | ResNet-50 interactive segmentation · Berkeley | NoC@90: 4.34 -> 4.36 | No clear effect |
| move the optimized scale and bias to before the second separable convolution block instead of the f-BRS-B location (ablation) | ResNet-50 interactive segmentation · Berkeley | NoC@90: 4.34 -> 4.91 | Hurt |
Closest papers by meaning
- CFR-ICL: Cascade-Forward Refinement with Iterative Click Loss for Interactive Image Segmentation (AAAI 2024 · similarity 0.83)
- FocSAM: Delving Deeply into Focused Objects in Segmenting Anything (CVPR 2024 · similarity 0.83)
- TETRIS: Towards Exploring the Robustness of Interactive Segmentation (AAAI 2024 · similarity 0.83)
- Adapt Before Comparison: A New Perspective on Cross-Domain Few-Shot Segmentation (CVPR 2024 · similarity 0.83)
- Dual-Agent Optimization framework for Cross-Domain Few-Shot Segmentation (CVPR 2025 · similarity 0.83)
- Rethinking Interactive Image Segmentation with Low Latency High Quality and Diverse Prompts (CVPR 2024 · 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.