Research map / Detection and segmentation / 3D detection and driving
MVSalNet:Multi-View Augmentation for RGB-D Salient Object Detection
Jiayuan Zhou, Lijun Wang, Huchuan Lu, Kaining Huang, Xinchu Shi, Bocong Liu
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 multi-view augmentation with three encoder-decoder branches (original plus two rotated views) (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.860 -> 0.901 | Improved |
| use novel views as extra training data only (ablation) | single encoder-decoder · NLPR | adaptive F-measure: 0.860 -> 0.874 | Improved |
| use rendered novel views instead of three branches with identical RGB (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.891 -> 0.901 | Improved |
| use two novel views instead of one (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.882 -> 0.901 | Improved |
| use rendered views instead of concatenating depth with RGB (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.895 -> 0.901 | Improved |
| add two more views at ±60 degrees (4 views) (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.901 -> 0.899 | No clear effect |
| change rotation angle from ±30 to ±60 degrees (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.901 -> 0.889 | Hurt |
| add Transformer dynamic filtering (TDF) instead of plain skip connection (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.886 -> 0.901 | Improved |
| use multi-dilation SW-MSA heads instead of single head (ablation) | TDF · NLPR | adaptive F-measure: 0.892 -> 0.901 | Improved |
| use TDF instead of CNN dynamic filtering (DDPM) (ablation) | MVSalNet · NLPR | adaptive F-measure: 389M -> 83M parameters; 0.896 -> 0.901 | Improved |
| use learned fusion module instead of adding the saliency maps (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.883 -> 0.901 | Improved |
| fuse into original view instead of supervising two synthesized maps separately (ablation) | MVSalNet · NLPR | adaptive F-measure: 0.896 -> 0.901 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- JL-DCF — JL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object Detection (CVPR 2020) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Promoting Saliency From Depth: Deep Unsupervised RGB-D Saliency Detection (ICLR 2022 · similarity 0.92)
- ViDSOD-100: A New Dataset and a Baseline Model for RGB-D Video Salient Object Detection (IJCV 2024 · similarity 0.91)
- Deep RGB-D Saliency Detection With Depth-Sensitive Attention and Automatic Multi-Modal Fusion (CVPR 2021 · similarity 0.91)
- Accurate RGB-D Salient Object Detection via Collaborative Learning (ECCV 2020 · similarity 0.91)
- Alignment-Free RGB-T Salient Object Detection: A Large-Scale Dataset and Progressive Correlation Network (AAAI 2025 · similarity 0.89)
- DiMSOD: A Diffusion-Based Framework for Multi-Modal Salient Object Detection (AAAI 2025 · 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.
