AI Paper Atlas

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

ECCV 2020 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Geometric alignment and synthetic-to-real adaptation · 137 citations (OpenAlex)

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 changedWhereMeasuredEffect
add finetuning of the generative model G2e with the YOLO classifier-feature cosine loss (ablation)YOLO-in-the-Dark · SID datasetmAP@IoU0.5: 0.37 -> 0.55Improved
use the YOLO-in-the-Dark model instead of original YOLO on brightness-enhanced RGBYOLO · SID datasetmAP@IoU0.5: 0.26 -> 0.55Improved
retrain YOLO on artificially darkened COCO instead of using SID front endYOLO · SID datasetmAP@IoU0.5: 0.26 -> 0.05Hurt
use naive SID+YOLO (full SID decoder) instead of glue layer mergingYOLO-in-the-Dark · SID datasetmAP@IoU0.5: 0.55 -> 0.56No clear effect
add gamma correction 0.67 to COCO training images (ablation)YOLO-in-the-Dark (no G2e finetuning) · SID datasetmAP@IoU0.5: 0.37 -> 0.36Hurt
add gamma correction 1.25 to COCO training images (ablation)YOLO-in-the-Dark (no G2e finetuning) · SID datasetmAP@IoU0.5: 0.37 -> 0.40Improved
add gamma correction 1.5 to COCO training images (ablation)YOLO-in-the-Dark (no G2e finetuning) · SID datasetmAP@IoU0.5: 0.37 -> 0.37No clear effect
add brightness and contrast data augmentation (ablation)YOLO-in-the-Dark (no G2e finetuning) · SID datasetmAP@IoU0.5: 0.37 -> 0.35Hurt
drop the high-resolution encoder scales (use fewer latent features) (ablation)SID latent feature reconstruction · SID reconstructionPSNR: 31.81 -> 25.92Hurt
replace SID decoder with glue layerSID+YOLO · 832x832 RAW imagetotal GMACs: 270.79 -> 113.90Improved
use YOLO-in-the-Dark instead of original YOLO at 0.013 luxYOLO · Self-collected indoor images at 0.013 luxminimum exposure time: 810 -> 333 msImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy