Research map / Detection and segmentation / Open-set and few-shot detection
DUNIT: Detection-Based Unsupervised Image-to-Image Translation
Check 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 |
|---|---|---|---|
| add instance-level feature processing and fuse it with global image features | DUNIT · INIT dataset, average | IS: 1.179 -> 1.223 | Helped |
| add instance-level feature processing and fuse it with global image features | DUNIT · INIT dataset, average | LPIPS distance: 0.274 -> 0.287 | Helped |
| translate source images with DUNIT before detection | object detection domain adaptation · VOC → Comics | mAP: 36.4 -> 40.2 | Helped |
| translate source images with DUNIT before detection | object detection domain adaptation · KITTI → Cityscapes | mAP: 46.1 -> 54.1 | Helped |
| add instance consistency loss (ablation) | DUNIT · Sunny → Night, INIT dataset | LPIPS distance: 0.322 -> 0.338 | Helped |
| add instance consistency loss (ablation) | DUNIT · Sunny → Rainy, INIT dataset | IS: 1.138 -> 1.225 | Helped |
| add instance consistency loss (ablation) | DUNIT · Sunny → Cloudy, INIT dataset | CIS: 1.009 -> 1.033 | Helped |
| add instance consistency loss (ablation) | DUNIT domain adaptation · KITTI → Cityscapes | mAP: 47.2 -> 54.1 | Helped |
| add instance consistency loss (ablation) | DUNIT domain adaptation · VOC → Comics | mAP: 39.4 -> 40.2 | Helped |
| use RetinaNet instead of Faster-RCNN (ablation) | DUNIT · Sunny → Night, INIT dataset | IS: 1.223 -> 1.259 | Helped |
| set RetinaNet α to 0.50 while keeping γ = 2 (ablation) | DUNIT · Sunny → Night, INIT dataset | IS: 1.259 -> 1.255 | No clear effect |
| set RetinaNet α = 1.0 and γ = 1.0 (ablation) | DUNIT · Sunny → Night, INIT dataset | IS: 1.259 -> 1.230 | Hurt |
Closest papers by meaning
- Towards Instance-Level Image-To-Image Translation (CVPR 2019 · 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 0.89)
- Unsupervised Image-to-Image Translation Networks (NeurIPS 2017 · similarity 0.88)
- Unsupervised Image-to-Image Translation Using Domain-Specific Variational Information Bound (NeurIPS 2018 · similarity 0.87)
- Memory-Guided Unsupervised Image-to-Image Translation (CVPR 2021 · similarity 0.87)
- Unsupervised Object Detection with Theoretical Guarantees (NeurIPS 2024 · similarity 0.86)
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.