Research map / Vision-language and multimodal / Human and embodied modeling
Explicit Visual Prompting for Low-Level Structure Segmentations
Weihuang Liu, Xi Shen, Chi-Man Pun, Xiaodong Cun
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 |
|---|---|---|---|
| Use EVP rather than task-specific detection methods | SegFormer-B4 low-level structure segmentation · DUT | Fβ: .854 -> .890 | Improved |
| Use EVP rather than task-specific detection methods | SegFormer-B4 low-level structure segmentation · SBU | BER: 3.04 -> 4.31 | Hurt |
| Use EVP rather than task-specific detection methods | SegFormer-B4 low-level structure segmentation · IMD20 | AUC: .848 -> .807 | Hurt |
| Use EVP rather than task-specific detection methods | SegFormer-B4 low-level structure segmentation · CAMO | Sα: .782 -> .846 | Improved |
| Use EVP with r=4 rather than full fine-tuning | SegFormer-B4 · ISTD | BER: 2.42 -> 1.35 | Improved |
| Add the patch embedding prompt F_pe (ablation) | EVP · CASIA | F1: .540 -> .636 | Improved |
| Add the high-frequency prompt F_hfc (ablation) | EVP · ISTD | BER: 1.61 -> 1.35 | Improved |
| Share MLP_tune across Adaptors (ablation) | EVP · CAMO | Sα: .846 -> .837 | Hurt |
| Use unshared MLP_up across Adaptors (ablation) | EVP · CASIA | F1: .636 -> .647 | Improved |
| Extend prompting from Stage 1 to Stages 1–4 (ablation) | EVP · CUHK | Fβ: .895 -> \bf.928 | Improved |
| Decrease scale factor from r=64 to r=4 (ablation) | EVP · CUHK | Fβ: .910 -> .928 | Improved |
| Decrease scale factor from r=4 to r=1 (ablation) | EVP · CUHK | Fβ: .928 -> \bf.931 | Improved |
Closest papers by meaning
- Visual Prompting for Generalized Few-shot Segmentation: A Multi-scale Approach (CVPR 2024 · similarity 0.83)
- Visual In-Context Prompting (CVPR 2024 · similarity 0.83)
- Rethinking Interactive Image Segmentation with Low Latency High Quality and Diverse Prompts (CVPR 2024 · similarity 0.82)
- Segment Anyword: Mask Prompt Inversion for Open-Set Grounded Segmentation (ICML 2025 · similarity 0.82)
- AutoVP: An Automated Visual Prompting Framework and Benchmark (ICLR 2024 · similarity 0.82)
- Curriculum Point Prompting for Weakly-Supervised Referring Image Segmentation (CVPR 2024 · similarity 0.82)
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.
