Research map / Detection and segmentation / CNN features
Learning Deep Relations to Promote Saliency Detection
Changrui Chen, Xin Sun, Yang Hua, Junyu Dong, Hongwei Xv
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 |
|---|---|---|---|
| refine PFA with MI relation model | PFA · ECSSD | F-measure: 0.8863 -> 0.9138 | Improved |
| refine weak conventional detector wCtrO | wCtrO · ECSSD | F-measure: 0.6763 -> 0.8537 | Improved |
| refine strongest detector R3Net | R3Net · ECSSD | F-measure: 0.9148 -> 0.9208 | Improved |
| refine PFA with MI relation model | PFA · ECSSD | MAE: 0.0448 -> 0.0383 | Improved |
| refine PiCANet | PiCANet · DUT-OMRON | F-measure: 0.7496 -> 0.7899 | Improved |
| use our method instead of LPS for refinement | Amulet · ECSSD | F-measure: 0.8963 -> 0.9069 | Improved |
| use our method instead of DenseCRFs | UCF · ECSSD | F-measure: 0.847 -> 0.878 | Improved |
| lower foreground threshold theta_fg from 0.9 to 0.7 at test (ablation) | SRM + ours · ECSSD | F-measure: 0.9158 -> 0.9152 | No clear effect |
| vary approximate thresholds theta_afg/theta_abg in training (ablation) | SRM + ours · ECSSD | F-measure: 0.9154 vs 0.9145 and 0.9149 | No clear effect |
Closest papers by meaning
- Revisiting Salient Object Detection: Simultaneous Detection, Ranking, and Subitizing of Multiple Salient Objects (CVPR 2018 · similarity 0.89)
- Non-Local Deep Features for Salient Object Detection (CVPR 2017 · similarity 0.88)
- Scene Context-Aware Salient Object Detection (ICCV 2021 · similarity 0.88)
- Salient Object Detection With Pyramid Attention and Salient Edges (CVPR 2019 · similarity 0.88)
- Instance-Level Relative Saliency Ranking with Graph Reasoning (TPAMI 2021 · similarity 0.88)
- Learning to Promote Saliency Detectors (CVPR 2018 · similarity 0.88)
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.
