Research map / Learning theory and algorithms / Prediction and decision losses
Spatially Binned ROC: A Comprehensive Saliency Metric
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 |
|---|---|---|---|
| evaluate fixation prediction separately in spatial bins using spROC | saliency algorithms · MIT and ImgSal | spatially resolved AUC | Mixed |
| weight bin scores by relative image area rather than fixation count | saliency algorithms, especially spatially biased models · MIT | AUC: GBVS 0.82 -> 0.78; CAS 0.78 -> 0.76 | Hurt |
| weight bin scores by relative image area rather than fixation count | saliency algorithms, especially spatially biased models · ImgSal | AUC: GBVS 0.74 -> 0.70; CAS 0.71 -> 0.69 | Hurt |
| add a small Gaussian smoothing kernel | AIM · MIT | AUC | Helped |
| increase Gaussian smoothing to very large kernels | AIM · MIT | peripheral-bin AUC | Hurt |
| use a centered Gaussian prior instead of image-content predictions | fixation prediction · MIT | AUC: 0.81 | Helped |
| use AWS rather than strongly spatially biased GBVS | saliency algorithm selection · MIT | spatial bias (standard deviation over bin AUC): GBVS 0.093; AWS 0.007 | Helped |
| use AWS rather than strongly spatially biased GBVS | saliency algorithm selection · ImgSal | spatial bias (standard deviation over bin AUC): GBVS 0.098; AWS 0.018 | Helped |
Closest papers by meaning
- Revisiting Saliency Metrics: Farthest-Neighbor Area Under Curve (CVPR 2020 · similarity 0.86)
- Advancing Saliency Ranking with Human Fixations: Dataset Models and Benchmarks (CVPR 2024 · similarity 0.86)
- What Do Deep Saliency Models Learn about Visual Attention? (NeurIPS 2023 · similarity 0.86)
- A Deeper Look at Saliency: Feature Contrast, Semantics, and Beyond (CVPR 2016 · similarity 0.86)
- A Motion-aware Spatio-temporal Graph for Video Salient Object Ranking (NeurIPS 2024 · similarity 0.85)
- Bridging the Divide: Reconsidering Softmax and Linear Attention (NeurIPS 2024 · 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.