Research map / Detection and segmentation / CNN features
A Bi-Directional Message Passing Model for Salient Object Detection
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 MCFEM with dilated convolutions to the FCN baseline (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.887 -> 0.914 | Helped |
| Replace MCFEM dilated convolutions with convolutions of kernel sizes 3, 7, 11, 15 (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.914 -> 0.909 | Hurt |
| Replace MCFEM with pyramid pooling (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.914 -> 0.894 | Hurt |
| Add bi-directional message passing instead of using a single directional stream (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.903 -> 0.917 | Helped |
| Add bi-directional message passing instead of using a single directional stream (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.894 -> 0.917 | Helped |
| Use adaptive gates rather than keeping all gates open (ablation) | GBMPM · ECSSD | max Fβ: 0.908 -> 0.917 | Helped |
| Use adaptive gates rather than closing all gates (ablation) | GBMPM · ECSSD | max Fβ: 0.889 -> 0.917 | Helped |
| Add GBMPM to the FCN baseline (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.887 -> 0.917 | Helped |
| Add both MCFEM and GBMPM to the FCN baseline (ablation) | VGG-16 salient object detector · ECSSD | max Fβ: 0.887 -> 0.928 | Helped |
| Use the proposed model rather than the strongest listed baseline | Salient object detection · ECSSD, PASCAL-S, SOD, HKU-IS, DUTS-test | max Fβ: 0.928 vs 0.917; 0.862 vs 0.847; 0.851 vs 0.841; 0.920 vs 0.910; 0.850 vs 0.827 | Helped |
| Use the proposed model rather than the strongest listed baseline | Salient object detection · ECSSD, PASCAL-S, SOD, HKU-IS, DUTS-test | MAE: 0.044 vs 0.052; 0.074 vs 0.085; 0.106 vs 0.118; 0.038 vs 0.041; 0.049 vs 0.057 | Helped |
Closest papers by meaning
- Deep Embedding Features for Salient Object Detection (AAAI 2019 · similarity 0.91)
- Progressive Attention Guided Recurrent Network for Salient Object Detection (CVPR 2018 · similarity 0.90)
- Weakly Supervised Salient Object Detection Using Image Labels (AAAI 2018 · similarity 0.90)
- Recurrently Aggregating Deep Features for Salient Object Detection (AAAI 2018 · similarity 0.90)
- An Iterative and Cooperative Top-Down and Bottom-Up Inference Network for Salient Object Detection (CVPR 2019 · similarity 0.90)
- A Mutual Learning Method for Salient Object Detection With Intertwined Multi-Supervision (CVPR 2019 · similarity 0.90)
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.