Research map / Detection and segmentation / Open-set and few-shot detection
PET-DINO: Unifying Visual Cues into Grounding DINO with Prompt-Enriched Training
Weifu Fu, Jinyang Li, Bin-Bin Gao, Jialin Li, Yuhuan Lin, Hanqiu Deng, Wenbing Tao, Yong Liu, Chengjie Wang
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 |
|---|---|---|---|
| replace T-Rex2 visual prompt encoder with AFVPG (ablation) | PET-DINO · COCO Visual-I (O365-trained) | AP: 59.2 -> 64.0 | Improved |
| replace T-Rex2 visual prompt encoder with AFVPG (ablation) | PET-DINO · COCO Visual-G (O365-trained) | AP: 37.6 -> 40.3 | Improved |
| add Intra-Batch Parallel Prompting to AFVPG (ablation) | PET-DINO · COCO Visual-G (O365-trained) | AP: 12.5 -> 37.2 | Improved |
| add Intra-Batch Parallel Prompting to AFVPG (ablation) | PET-DINO · COCO Visual-I (O365-trained) | AP: 67.0 -> 63.2 | Hurt |
| add Dynamic Memory-Driven Prompting to AFVPG (ablation) | PET-DINO · COCO Visual-G (O365-trained) | AP: 12.5 -> 24.7 | Improved |
| add DMD on top of AFVPG + IBP (ablation) | PET-DINO · COCO Visual-G (O365-trained) | AP: 37.2 -> 40.3 | Improved |
| inherit pre-trained MM-Grounding-DINO instead of training from scratch (ablation) | PET-DINO-T · COCO Visual-G (O365-trained) | AP: 32.7 -> 40.3 | Improved |
| inherit pre-trained MM-Grounding-DINO instead of training from scratch (ablation) | PET-DINO-T · COCO Visual-I (O365-trained) | AP: 62.1 -> 64.0 | Improved |
| add visual prompt route training to text-prompted detector (Swin-L) | MM-Grounding-DINO · COCO text prompt | AP: 53.0 -> 54.0 | Improved |
| add visual prompt route training to text-prompted detector (Swin-T) | MM-Grounding-DINO · COCO text prompt | AP: 50.6 -> 49.8 | Hurt |
| train on more data (O365, OpenImages, V3Det vs O365 only) | PET-DINO Swin-T · ODinW35 Visual-I | AP: 38.8 -> 48.3 | Improved |
| train on more data (O365, OpenImages, V3Det vs O365 only) | PET-DINO Swin-T · COCO Visual-G | AP: 40.3 -> 38.4 | Hurt |
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.
- T-Rex2 — T-Rex2: Towards Generic Object Detection via Text-Visual Prompt Synergy (ECCV 2024) · better on 5 of 7 shared results
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
- Just a Few Glances: Open-Set Visual Perception with Image Prompt Paradigm (AAAI 2025 · similarity 0.92)
- DoGA: Enhancing Grounded Object Detection via Grouped Pre-Training with Attributes (AAAI 2025 · similarity 0.89)
- CP-DETR: Concept Prompt Guide DETR Toward Stronger Universal Object Detection (AAAI 2025 · similarity 0.88)
- Visual Diversity and Region-aware Prompt Learning for Zero-shot HOI Detection (NeurIPS 2025 · similarity 0.88)
- DiPEx: Dispersing Prompt Expansion for Class-Agnostic Object Detection (NeurIPS 2024 · similarity 0.88)
- Learning Background Prompts to Discover Implicit Knowledge for Open Vocabulary Object Detection (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.
