Research map / Vision-language and multimodal / CLIP and zero-shot
Fine-Grained Visual Prompting
Lingfeng Yang, Yueze Wang, Xiang Li, Xinlong Wang, Jian Yang
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 |
|---|---|---|---|
| blur background outside mask (reverse mask) instead of colored mask (ablation) | FGVP · RefCOCO (REC proposals) | accuracy: 41.8 -> 52.8 | Improved |
| blur reverse instead of grayscale reverse mask (ablation) | FGVP · RefCOCOg (REC proposals) | accuracy: 51.2 -> 57.8 | Improved |
| use fine-grained mask instead of box for blur reverse (ablation) | FGVP · RefCOCO+ (REC proposals) | accuracy: 51.4 -> 55.4 | Improved |
| use mask instead of circle for blur reverse (ablation) | FGVP · RefCOCO (GT annotations) | accuracy: 49.2 -> 52.8 | Improved |
| crop instead of blur reverse mask (ablation) | Crop · COCO (GT annotations) | accuracy: 70.9 -> 67.8 | Hurt |
| crop instead of blur reverse mask (ablation) | Crop · RefCOCO (GT annotations) | accuracy: 35.2 -> 52.8 | Improved |
| colored mask without blur reverse (ablation) | Mask D2 · RefCOCO (GT annotations) | accuracy: 52.8 -> 29.9 | Hurt |
| Gaussian blur standard deviation set to 100 (ablation) | Blur Reverse Mask · 8 benchmarks | accuracy | Improved |
| replace RedCircle with FGVP, ViT-L, UNINEXT proposals | FGVP · RefCOCO val | accuracy: 48.9 -> 52.8 | Improved |
| replace RedCircle with FGVP in ReCLIP codebase with R+S | FGVP · RefCOCO val | accuracy: 56.8 -> 59.6 | Improved |
| replace ReCLIP prompts with FGVP | FGVP · RefCOCO+ testA | accuracy: 50.1 -> 60.4 | Improved |
| SAM scale base to huge without mask filter (ablation) | FGVP · RefCOCO | accuracy: 51.7 -> 51.9 | No clear effect |
Closest papers by meaning
- Visual In-Context Prompting (CVPR 2024 · similarity 0.92)
- Just a Few Glances: Open-Set Visual Perception with Image Prompt Paradigm (AAAI 2025 · similarity 0.88)
- LaViP: Language-Grounded Visual Prompting (AAAI 2024 · similarity 0.88)
- STOP: Integrated Spatial-Temporal Dynamic Prompting for Video Understanding (CVPR 2025 · similarity 0.88)
- Visual Prompting for Generalized Few-shot Segmentation: A Multi-scale Approach (CVPR 2024 · similarity 0.88)
- Tune-An-Ellipse: CLIP Has Potential to Find What You Want (CVPR 2024 · 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.
