Research map / Detection and segmentation / CNN features
Hierarchical Image Saliency Detection on Extended CSSD
Jianping Shi, Qiong Yan, Xu Li, Jiaya Jia
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 |
|---|---|---|---|
| aggregate three layers with local consistent hierarchical inference instead of the average of the layers (ablation) | CHS · ECSSD | F-measure: 0.6739 -> 0.6776 | Improved |
| use multi-layer CHS instead of the best single layer (Layer1) (ablation) | CHS · ECSSD | F-measure: 0.6489 -> 0.6776 | Improved |
| average the three layers instead of using a single layer (ablation) | Layer averaging · ECSSD | F-measure: 0.6440 -> 0.6739 | Improved |
| add local consistency term to tree-structured inference (HS to CHS) (ablation) | CHS · ECSSD | F-measure: 0.6721 -> 0.6776 | Improved |
| add local consistency term to tree-structured inference (HS to CHS) (ablation) | CHS · ECSSD | precision: 0.7360 -> 0.7639 | Improved |
| add local consistency term to tree-structured inference (HS to CHS) (ablation) | CHS · ECSSD | recall: 0.6832 -> 0.6564 | Hurt |
| add local consistency term to tree-structured inference (HS to CHS) (ablation) | CHS · ECSSD | MAE: 0.2244 -> 0.2265 | Hurt |
| add local consistency term to HS (ablation) | CHS · MSRA-1000 | MAE: 0.1155 -> 0.0961 | Improved |
| add local consistency term to HS (ablation) | CHS · MSRA-5000 | MAE: 0.1528 -> 0.1499 | No clear effect |
| use CHS instead of RCC | CHS · ECSSD | MAE: 0.1865 -> 0.2265 | Hurt |
| use CHS instead of RCC | CHS · MSRA-1000 | MAE: 0.1062 -> 0.0961 | Improved |
| replace pixel-count region scale with the proposed uniformity-based scale measure (ablation) | CHS · ECSSD | F-measure | Improved |
Closest papers by meaning
- Saliency Hierarchy Modeling via Generative Kernels for Salient Object Detection (ECCV 2022 · similarity 0.89)
- ELD-Net: An Efficient Deep Learning Architecture for Accurate Saliency Detection (TPAMI 2017 · similarity 0.88)
- Non-Local Deep Features for Salient Object Detection (CVPR 2017 · similarity 0.88)
- Ranking Saliency (TPAMI 2016 · similarity 0.88)
- GraB: Visual Saliency via Novel Graph Model and Background Priors (CVPR 2016 · similarity 0.88)
- Saliency Pattern Detection by Ranking Structured Trees (ICCV 2017 · 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.
