Research map / Architectures and efficiency / Attention and transformers
VST++: Efficient and Stronger Visual Saliency Transformer
Nian Liu, Ziyang Luo, Ni Zhang, Junwei Han
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 token-supervised prediction loss (ablation) | VST · ECSSD | MAE: 0.034 -> 0.029 | Improved |
| add token-supervised prediction loss (ablation) | VST · ECSSD | MACs (G): 41.36 -> 42.67 | Hurt |
| add token-supervised prediction loss (ablation) | VST · SSD (RGB-D) | MAE: 0.045 -> 0.038 | Improved |
| add token-supervised prediction loss (ablation) | VST · DUT-O | maxF: 0.800 -> 0.810 | Improved |
| add select-integrate attention (ablation) | VST + L_sup · RGB-D benchmarks | MACs (G): 52.64 -> 39.98 | Improved |
| add select-integrate attention (ablation) | VST + L_sup · RGB SOD | MACs (G): 42.67 -> 30.07 | Improved |
| add select-integrate attention (ablation) | VST + L_sup · ECSSD | S-measure: 0.935 -> 0.933 | No clear effect |
| add select-integrate attention (ablation) | VST + L_sup · SSD (RGB-D) | maxF: 0.884 -> 0.873 | Hurt |
| add select-integrate attention (ablation) | VST + L_sup · NLPR | S-measure: 0.934 -> 0.931 | No clear effect |
| add depth position encoding (ablation) | VST + L_sup + SIA · NLPR | MAE: 0.023 -> 0.020 | Improved |
| add depth position encoding (ablation) | VST + L_sup + SIA · ReDWeb-S | MAE: 0.108 -> 0.103 | Improved |
| add depth position encoding (ablation) | VST + L_sup + SIA · RGB-D benchmarks | MACs (G): 39.98 -> 39.95 | No clear effect |
Closest papers by meaning
- Visual Saliency Transformer (ICCV 2021 · similarity 0.96)
- A2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object Detection (CVPR 2020 · similarity 0.88)
- DiMSOD: A Diffusion-Based Framework for Multi-Modal Salient Object Detection (AAAI 2025 · similarity 0.87)
- You Only Need Less Attention at Each Stage in Vision Transformers (CVPR 2024 · similarity 0.87)
- Alignment-Free RGB-T Salient Object Detection: A Large-Scale Dataset and Progressive Correlation Network (AAAI 2025 · similarity 0.87)
- Attention Guided CAM: Visual Explanations of Vision Transformer Guided by Self-Attention (AAAI 2024 · 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.
