Research map / Detection and segmentation / CNN features
Asymmetric Two-Stream Architecture for Accurate RGB-D Saliency Detection
Miao Zhang, Sun Xiao Fei, Jie Liu, Shuang Xu, Yongri Piao, Huchuan Lu
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 |
|---|---|---|---|
| add flow ladder module (FLM) to both streams (ablation) | symmetric two-stream VGG-19 baseline · DUT-RGBD | F-measure: 0.822 -> 0.911 | Improved |
| add FLM (ablation) | symmetric baseline · NLPR | F-measure: 0.766 -> 0.872 | Improved |
| replace summation fusion with depth attention module (DAM) (ablation) | symmetric baseline · NLPR | MAE: 0.050 -> 0.041 | Improved |
| replace summation fusion with DAM (ablation) | symmetric baseline · NJUD | F-measure: 0.810 -> 0.811 | No clear effect |
| replace summation fusion with DAM (ablation) | symmetric baseline · SSD | F-measure: 0.749 -> 0.738 | Hurt |
| add DAM to asymmetric FLM model (ablation) | B+FLM[a] · DUT-RGBD | F-measure: 0.909 -> 0.920 | Improved |
| add DAM to asymmetric FLM model (ablation) | B+FLM[a] · NLPR | F-measure: 0.870 -> 0.876 | Improved |
| replace VGG-19 depth stream with lightweight DepthNet (ablation) | B+FLM · DUT-RGBD | F-measure: 0.911 -> 0.909 | No clear effect |
| replace symmetric depth stream with DepthNet (model size) (ablation) | B+FLM+DAM · model size | MB: 244.4MB -> 128.9MB | Improved |
| use asymmetric instead of symmetric streams with FLM+DAM (ablation) | B+FLM+DAM · NLPR | F-measure: 0.890 -> 0.876 | Hurt |
| use asymmetric instead of symmetric streams with FLM+DAM (ablation) | B+FLM+DAM · LFSD | F-measure: 0.863 -> 0.862 | No clear effect |
| add FLM with asymmetric depth stream (ablation) | asymmetric baseline · LFSD | F-measure: 0.812 -> 0.870 | Improved |
Closest papers by meaning
- A2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object Detection (CVPR 2020 · similarity 0.92)
- A Single Stream Network for Robust and Real-time RGB-D Salient Object Detection (ECCV 2020 · similarity 0.92)
- Select, Supplement and Focus for RGB-D Saliency Detection (CVPR 2020 · similarity 0.90)
- Promoting Saliency From Depth: Deep Unsupervised RGB-D Saliency Detection (ICLR 2022 · similarity 0.90)
- RGB-D Saliency Detection via Cascaded Mutual Information Minimization (ICCV 2021 · similarity 0.90)
- Exploit and Replace: An Asymmetrical Two-Stream Architecture for Versatile Light Field Saliency Detection (AAAI 2020 · 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.
