Research map / Detection and segmentation / Open-set and few-shot detection
Cross-Domain Few-Shot Object Detection via Enhanced Open-Set Object Detector
Yuqian Fu, Yu Wang, Yixuan Pan, Xingyu Qiu, Lian Huai, Zeyu Shangguan, Tong Liu, Yanwei Fu, Luc Van Gool, Xingqun Jiang
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 |
|---|---|---|---|
| finetune the top detection and classification heads of DE-ViT (ablation) | DE-ViT · CD-FSOD benchmark (6 targets avg, 10-shot) | mAP: 5.9 -> 25.2 | Improved |
| add learnable instance features, reweighting, and domain prompter on top of finetuned DE-ViT (ablation) | DE-ViT-FT · CD-FSOD benchmark (6 targets avg, 10-shot) | mAP: 25.2 -> 29.6 | Improved |
| add learnable instance features (ablation) | DE-ViT-FT · CD-FSOD benchmark (10-shot) | mAP | Improved |
| add instance reweighting and domain prompter modules (10-shot) (ablation) | CD-ViTO · CD-FSOD benchmark (10-shot) | mAP | Improved |
| add the domain prompter in the 1-shot setting (ablation) | CD-ViTO · CD-FSOD benchmark (1-shot) | mAP | Mixed |
| use open-set detector without finetuning on cross-domain targets | DE-ViT · CD-FSOD benchmark (6 targets avg, 1-shot) | mAP: 1.0 vs 10.1 (DE-ViT-FT) | Hurt |
| use ViT backbone instead of ResNet50 for CD-FSOD | FSOD detectors · CD-FSOD benchmark | mAP | Mixed |
| learnable instance features enlarge inter-class distances of prototypes (ablation) | CD-ViTO · NEU-DET (5-shot) | cosine distance between class prototypes | Improved |
| significant indefinable boundaries in the target dataset | All evaluated detectors · NEU-DET and UODD | mAP: CD-ViTO 12.8 and 7.0 at 10-shot | Hurt |
Closest papers by meaning
- Less Is Better: Sparse Instance Learning for Cross-Domain Few-Shot Object Detection (AAAI 2026 · similarity 0.92)
- OA-FSUI2IT: A Novel Few-Shot Cross Domain Object Detection Framework with Object-Aware Few-Shot Unsupervised Image-to-Image Translation (AAAI 2022 · similarity 0.91)
- Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection (NeurIPS 2025 · similarity 0.91)
- Revisiting Few-Shot Object Detection with Vision-Language Models (NeurIPS 2024 · similarity 0.89)
- Fine-Grained Prototypes Distillation for Few-Shot Object Detection (AAAI 2024 · similarity 0.89)
- Improving Single Domain-Generalized Object Detection: A Focus on Diversification and Alignment (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.
