Research map / Label-efficient and robust learning / Domain adaptation and generalization
YOLO in the Dark - Domain Adaptation Method for Merging Multiple Models -
Yukihiro Sasagawa, Hajime Nagahara
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 finetuning of the generative model G2e with the YOLO classifier-feature cosine loss (ablation) | YOLO-in-the-Dark · SID dataset | mAP@IoU0.5: 0.37 -> 0.55 | Improved |
| use the YOLO-in-the-Dark model instead of original YOLO on brightness-enhanced RGB | YOLO · SID dataset | mAP@IoU0.5: 0.26 -> 0.55 | Improved |
| retrain YOLO on artificially darkened COCO instead of using SID front end | YOLO · SID dataset | mAP@IoU0.5: 0.26 -> 0.05 | Hurt |
| use naive SID+YOLO (full SID decoder) instead of glue layer merging | YOLO-in-the-Dark · SID dataset | mAP@IoU0.5: 0.55 -> 0.56 | No clear effect |
| add gamma correction 0.67 to COCO training images (ablation) | YOLO-in-the-Dark (no G2e finetuning) · SID dataset | mAP@IoU0.5: 0.37 -> 0.36 | Hurt |
| add gamma correction 1.25 to COCO training images (ablation) | YOLO-in-the-Dark (no G2e finetuning) · SID dataset | mAP@IoU0.5: 0.37 -> 0.40 | Improved |
| add gamma correction 1.5 to COCO training images (ablation) | YOLO-in-the-Dark (no G2e finetuning) · SID dataset | mAP@IoU0.5: 0.37 -> 0.37 | No clear effect |
| add brightness and contrast data augmentation (ablation) | YOLO-in-the-Dark (no G2e finetuning) · SID dataset | mAP@IoU0.5: 0.37 -> 0.35 | Hurt |
| drop the high-resolution encoder scales (use fewer latent features) (ablation) | SID latent feature reconstruction · SID reconstruction | PSNR: 31.81 -> 25.92 | Hurt |
| replace SID decoder with glue layer | SID+YOLO · 832x832 RAW image | total GMACs: 270.79 -> 113.90 | Improved |
| use YOLO-in-the-Dark instead of original YOLO at 0.013 lux | YOLO · Self-collected indoor images at 0.013 lux | minimum exposure time: 810 -> 333 ms | Improved |
Closest papers by meaning
- DLDA: Unified Dual-Level Domain Adaptation for Low-Light Object Detection (AAAI 2026 · similarity 0.87)
- Diversify and Match: A Domain Adaptive Representation Learning Paradigm for Object Detection (CVPR 2019 · similarity 0.86)
- YOLO-RD: Introducing Relevant and Compact Explicit Knowledge to YOLO by Retriever-Dictionary (ICLR 2025 · similarity 0.85)
- Boosting Object Detection with Zero-Shot Day-Night Domain Adaptation (CVPR 2024 · similarity 0.85)
- You Only Look Around: Learning Illumination-Invariant Feature for Low-light Object Detection (NeurIPS 2024 · similarity 0.85)
- Large Self-Supervised Models Bridge the Gap in Domain Adaptive Object Detection (CVPR 2025 · similarity 0.84)
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.
