AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Revisiting Attention in the Dark for Low-Light Person Re-Identiffcation

Xiang Guo, Ruimin Hu, Dongliang Zhu, Mei Wang

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Supervised person re-identification · 0 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 Norm-Ratio Attention to baseline (ablation)TransReID-SSL ViT-B/16+ICS · Night600mAP (%): 9.5 -> 10.4Improved
add NRA on top of multi-domain learning (ablation)TransReID-SSL ViT-B/16+ICS · Night600Rank-1 (%): 24.2 -> 27.1Improved
add NRA on top of SRD (ablation)TransReID-SSL ViT-B/16+ICS · Night600mAP (%): 11.4 -> 12.5Improved
use SRD instead of standard multi-domain learning (ablation)TransReID-SSL ViT-B/16+ICS · Night600Rank-1 (%): 24.2 -> 25.3Improved
add NRA to ViT-S/16 without ICS (ablation)TransReID-SSL ViT-S/16 · Night600mAP (%): 8.5 -> 8.2Hurt
add NRA to ViT-S/16 with ICS (ablation)TransReID-SSL ViT-S/16+ICS · Night600mAP (%): 8.3 -> 8.7Improved
add NRA to ViT-B/16 with ICS (ablation)TransReID-SSL ViT-B/16+ICS · Night600Rank-5 (%): 36.1 -> 39.0Improved
add NRA under normal lighting (ablation)TransReID-SSL · Market1501mAP (%): 93.1 -> 93.1No clear effect
add NRA under normal lighting (ablation)TransReID-SSL · MSMT17mAP (%): 75.1 -> 73.4Hurt
NRSRD vs CNet on RGBNT201rgbNRSRD · RGBNT201rgbmAP (%): 55.3 -> 58.9Improved
NRSRD vs CNet on Night600NRSRD · Night600mAP (%): 13.3 -> 12.5Hurt

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