Research map / Detection and segmentation / Open-set and few-shot detection
Focused and Collaborative Feedback Integration for Interactive Image Segmentation
Read the paperCheck 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 |
|---|---|---|---|
| add FFCM and CFFM (ablation) | FCFI with ResNet-101 · Berkeley and DAVIS | NoC@90%: 4.31 -> 2.84; 7.56 -> 6.48 | Helped |
| add FFCM and CFFM (ablation) | FCFI with HRNet-18 · Berkeley and DAVIS | NoC@90%: 2.60 -> 1.96; 5.73 -> 5.16 | Helped |
| add FFCM (ablation) | baseline with ResNet-101 · Berkeley and DAVIS | NoC@90%: 4.31 -> 3.75; 7.56 -> 6.75 | Helped |
| add CFFM (ablation) | baseline with ResNet-101 · Berkeley and DAVIS | NoC@90%: 4.31 -> 3.07; 7.56 -> 6.62 | Helped |
| use cosine rather than exponential feature similarity (ablation) | FFCM · Berkeley and DAVIS | NoC@90%: 3.60 -> 2.85; 7.51 -> 5.82 | Helped |
| set expansion ratio r to 0.3 rather than 0.1, 0.2, 0.4, or 0.5 (ablation) | FFCM · Berkeley and DAVIS | NoC@90%: 3.43, 3.10, 2.85, 3.12, 3.27; 7.46, 6.82, 5.82, 6.42, 6.72 | Mixed |
| fuse feedback into deep features rather than concatenate it with the input (ablation) | CFFM · Berkeley and DAVIS | NoC@90%: 4.00 -> 3.07; 7.08 -> 6.10 | Helped |
| retain the residual connection (ablation) | CFFM · Berkeley and DAVIS | NoC@90%: 3.52 -> 3.07; 6.63 -> 6.10 | Helped |
| retain the interaction gate (ablation) | CFFM · Berkeley and DAVIS | NoC@90%: 3.36 -> 3.07; 6.37 -> 6.10 | Helped |
| integrate previous-round feedback into deep features (ablation) | FCFI · Berkeley and DAVIS | NoC@90%: 4.31 -> 3.07; 7.56 -> 6.10 | Helped |
| use FCFI rather than RITM | HRNet-18 · GrabCut, Berkeley, SBD, and DAVIS | NoC@90%: 1.54 -> 1.46; 2.26 -> 1.96; 6.06 -> 5.83; 5.74 -> 5.16 | Helped |
| use FCFI rather than FocusCut | ResNet-101 · Berkeley and DAVIS | NoC@90%: 3.01 -> 2.84; 6.22 -> 6.48 | Mixed |
Closest papers by meaning
- CFR-ICL: Cascade-Forward Refinement with Iterative Click Loss for Interactive Image Segmentation (AAAI 2024 · similarity 0.88)
- TETRIS: Towards Exploring the Robustness of Interactive Segmentation (AAAI 2024 · similarity 0.87)
- Rethinking Interactive Image Segmentation with Low Latency High Quality and Diverse Prompts (CVPR 2024 · similarity 0.86)
- RClicks: Realistic Click Simulation for Benchmarking Interactive Segmentation (NeurIPS 2024 · similarity 0.86)
- MFP: Making Full Use of Probability Maps for Interactive Image Segmentation (CVPR 2024 · similarity 0.86)
- FocSAM: Delving Deeply into Focused Objects in Segmenting Anything (CVPR 2024 · similarity 0.86)
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.