Research map / Detection and segmentation / Open-set and few-shot detection
Structure-Measure: A New Way to Evaluate Foreground Maps
Deng-Ping Fan, Ming-Ming Cheng, Yun Liu, Tao Li, Ali Borji
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 |
|---|---|---|---|
| use Structure-measure instead of Fbw | evaluation measure · PASCAL-S (MM1) | 1-Spearman rho: 0.365 -> 0.320 | Improved |
| use Structure-measure instead of Fbw | evaluation measure · HKU-IS (MM1) | 1-Spearman rho: 0.498 -> 0.424 | Improved |
| use Structure-measure instead of Fbw | evaluation measure · PASCAL-S (MM3) | GT-switch error (%): 1.05 -> 0.34 | Improved |
| use Structure-measure instead of Fbw | evaluation measure · HKU-IS (MM3) | GT-switch error (%): 0.26 -> 0.08 | Improved |
| use Structure-measure instead of Fbw | evaluation measure · ECSSD (MM2) | SOTA-vs-generic error (%): 3.00 -> 3.30 | Hurt |
| use Structure-measure instead of Fbw | evaluation measure · SOD (MM2) | SOTA-vs-generic error (%): 16.3 -> 9.67 | Improved |
| use Structure-measure instead of AUC | evaluation measure · PASCAL-S (MM2) | SOTA-vs-generic error (%): 15.8 -> 4.59 | Improved |
| use Structure-measure instead of AP/AUC/Fbw | evaluation measure · ASD1000 (MM2) | generic-map wins over 1000 images: 11 errors | Improved |
| use Structure-measure instead of Fbw (annotation-error robustness) | evaluation measure · ASD1000 (MM4) | 1-Spearman rho | No clear effect |
| choose map by Structure-measure instead of Fbw (user study) | evaluation measure · user study (100 map pairs) | preference rate (%): 63.69% | Improved |
| choose map by Structure-measure instead of AP (user study) | evaluation measure · user study | preference rate (%): 72.11% | Improved |
| choose map by Structure-measure instead of AUC (user study) | evaluation measure · user study | preference rate (%): 73.56% | Improved |
Closest papers by meaning
- FOCUS: Towards Universal Foreground Segmentation (AAAI 2025 · similarity 0.86)
- Training Deep Networks to Be Spatially Sensitive (ICCV 2017 · similarity 0.86)
- Towards Size-invariant Salient Object Detection: A Generic Evaluation and Optimization Approach (TPAMI 2025 · similarity 0.85)
- Ranking Saliency (TPAMI 2016 · similarity 0.85)
- T-measure: A Topology-Consistent Metric for Binary Segmentation (ICML 2026 · similarity 0.85)
- Salient Objects in Clutter: Bringing Salient Object Detection to the Foreground (ECCV 2018 · 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.
