Research map / Detection and segmentation / CNN features
Learning Selective Self-Mutual Attention for RGB-D Saliency Detection
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 |
|---|---|---|---|
| add DenseASPP to the baseline UNet (ablation) | RGB-D saliency UNet · NJUD | maxF: 0.852 -> 0.865 | Helped |
| add Non-local attention to UNet + DenseASPP (ablation) | RGB-D saliency UNet · NJUD | maxF: 0.865 -> 0.865 | No clear effect |
| replace Non-local attention with self-mutual attention (ablation) | RGB-D saliency UNet + DenseASPP · NJUD | maxF: 0.865 -> 0.882 | Helped |
| add selection attention to self-mutual attention (ablation) | RGB-D saliency UNet + DenseASPP · NJUD | maxF: 0.882 -> 0.884 | Helped |
| add residual fusion of depth decoder features into the RGB stream (ablation) | S2MA saliency model · NJUD | maxF: 0.884 -> 0.889 | Helped |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · RGBD135 | maxF: 0.888 -> 0.935 | Helped |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · LFSD | maxF: 0.856 -> 0.835 | Hurt |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · NJUD | MAE: 0.051 -> 0.053 | Hurt |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · NLPR | MAE: 0.031 -> 0.030 | Helped |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · STERE | maxF: 0.886 -> 0.882 | Hurt |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · SSD | maxF: 0.844 -> 0.848 | Helped |
| use S2MA instead of the strongest baseline | RGB-D saliency detection · DUT-RGBD | maxF: 0.898 -> 0.901 | Helped |
Closest papers by meaning
- Select, Supplement and Focus for RGB-D Saliency Detection (CVPR 2020 · similarity 0.90)
- Deep RGB-D Saliency Detection With Depth-Sensitive Attention and Automatic Multi-Modal Fusion (CVPR 2021 · similarity 0.90)
- Promoting Saliency From Depth: Deep Unsupervised RGB-D Saliency Detection (ICLR 2022 · similarity 0.90)
- JL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object Detection (CVPR 2020 · similarity 0.89)
- UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders (CVPR 2020 · similarity 0.88)
- A2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object Detection (CVPR 2020 · similarity 0.87)
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.