Research map / Detection and segmentation / Open-set and few-shot detection
Visual Textualization for Image Prompted Object Detection
Yongjian Wu, Yang Zhou, Jiya Saiyin, Bingzheng Wei, Yan Xu
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 visual textualization (VisTex) to GLIP | GLIP · PASCAL VOC novel split | mean AP50: 54.2 -> 71.8 | Improved |
| add VisTex to GroundingDINO | GroundingDINO · LVIS MiniVal | AP: 38.3 -> 42.8 | Improved |
| add VisTex to GLIP | GLIP · ODinW13 2-shot | mean AP: 38.5 -> 45.1 | Improved |
| add VisTex to RegionCLIP | RegionCLIP · COCO open-vocabulary | novel AP50: 61.1 -> 65.3 | Improved |
| add text prompt to image prompt (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 30.6 -> 37.1 | Improved |
| add MSTB parameter sharing (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 37.1 -> 48.3 | Improved |
| add multi-stage fusion (MSF) (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 48.3 -> 50.3 | Improved |
| use stages 1-8 instead of only stages 5-8 (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 23.2 -> 50.3 | Improved |
| use stages 1-8 instead of only stage 1 (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 48.5 -> 50.3 | Improved |
| use multi-scale features (0+1+2+3+4) with sharing instead of scale 0 only (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 47.1 -> 50.3 | Improved |
| fuse shots by concatenation instead of element-wise addition (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 41.6 -> 50.3 | Improved |
| fuse stages by max pooling instead of concatenation (ablation) | VisTex-GLIP · MSCOCO 2-shot | AP: 39.4 -> 50.3 | Improved |
Closest papers by meaning
- LLMs Meet VLMs: Boost Open Vocabulary Object Detection with Fine-grained Descriptors (ICLR 2024 · similarity 0.87)
- Simple Image-Level Classification Improves Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.87)
- Just a Few Glances: Open-Set Visual Perception with Image Prompt Paradigm (AAAI 2025 · similarity 0.86)
- Roboflow100-VL: A Multi-Domain Object Detection Benchmark for Vision-Language Models (NeurIPS 2025 · similarity 0.86)
- VisMin: Visual Minimal-Change Understanding (NeurIPS 2024 · similarity 0.85)
- DocVLM: Make Your VLM an Efficient Reader (CVPR 2025 · similarity 0.85)
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.
