Research map / Detection and segmentation / Open-set and few-shot detection
OA-FSUI2IT: A Novel Few-Shot Cross Domain Object Detection Framework with Object-Aware Few-Shot Unsupervised Image-to-Image Translation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 OA-FSUI2IT instead of CUT for few-shot translation and detection | Faster R-CNN · Normal-to-Foggy, ten-shot | mAP: 39.5 -> 42.3 | Helped |
| Use OA-FSUI2IT instead of CUT for few-shot translation and detection | Faster R-CNN · Normal-to-Foggy, one-shot | mAP: 34.6 -> 40.8 | Helped |
| Use OA-FSUI2IT instead of CUT for few-shot translation and detection | Faster R-CNN · Cross-scene adaptation (Kitti → Cityscapes) | mAP: 21.7 -> 26.3 | Helped |
| Use OA-FSUI2IT instead of CUT for few-shot translation and detection | Faster R-CNN · Day-to-Night (BDD100k) | mAP: 26.3 -> 30.5 | Helped |
| Add discriminator augmentation (ablation) | OA-FSUI2IT · Normal-to-Foggy | mAP: 39.5 -> 41.1 | Helped |
| Add patch pyramid contrastive learning (ablation) | OA-FSUI2IT · Normal-to-Foggy | mAP: 39.5 -> 41.0 | Helped |
| Add self-supervised content-consistency loss alone (ablation) | OA-FSUI2IT · Normal-to-Foggy | mAP: 39.5 -> 39.0 | Hurt |
| Add DA and PPCL together, without SSCC (ablation) | OA-FSUI2IT · Normal-to-Foggy | mAP: 39.5 -> 41.3 | Helped |
| Add DA, PPCL, and SSCC together (ablation) | OA-FSUI2IT · Normal-to-Foggy | mAP: 39.5 -> 42.3 | Helped |
| Use the PPCL variant shown with 9 instead of the variant shown with 6 (ablation) | OA-FSUI2IT without DA or SSCC · Normal-to-Foggy | mAP: 41.0 -> 41.7 | Helped |
| Use the PPCL variant shown with 9 alongside DA and SSCC (ablation) | OA-FSUI2IT · Normal-to-Foggy | mAP: 42.3 -> 42.3 | No clear effect |
Closest papers by meaning
- DUNIT: Detection-Based Unsupervised Image-to-Image Translation (CVPR 2020 · similarity 0.89)
- DON’T NEED RETRAINING: A Mixture of DETR and Vision Foundation Models for Cross-Domain Few-Shot Object Detection (NeurIPS 2025 · similarity 0.89)
- FSCE: Few-Shot Object Detection via Contrastive Proposal Encoding (CVPR 2021 · similarity 0.89)
- Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection (NeurIPS 2025 · similarity 0.88)
- Object-Aware Domain Generalization for Object Detection (AAAI 2024 · similarity 0.88)
- Open-Det: An Efficient Learning Framework for Open-Ended Detection (ICML 2025 · similarity 0.87)
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.