AI Paper Atlas

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

Modality Unifying Network for Visible-Infrared Person Re-Identification

Hao Yu, Xu Cheng, Wei Peng, Weihao Liu, Guoying Zhao

ICCV 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Cross-modality and occlusion re-identification · 111 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 auxiliary modality generated by IML and CML (ablation)ResNet-50 baseline · SYSU-MM01 all-searchrank-1 (%): 57.49 -> 62.55Improved
add identity consistency loss L_idc (ablation)baseline + Aux. · SYSU-MM01 all-searchrank-1 (%): 62.55 -> 66.58Improved
add identity alignment loss L_ia (ablation)baseline + Aux. + L_idc · SYSU-MM01 all-searchrank-1 (%): 66.58 -> 71.35Improved
add modality alignment loss L_ma (ablation)baseline + Aux. + L_idc · SYSU-MM01 all-searchrank-1 (%): 66.58 -> 69.77Improved
add both L_ia and L_ma (ablation)baseline + Aux. + L_idc · SYSU-MM01 all-searchrank-1 (%): 66.58 -> 76.24Improved
use the proposed auxiliary modality instead of syncretic modality (ablation)MUN · SYSU-MM01 all-searchrank-1 (%): 72.95 -> 76.24Improved
use the proposed auxiliary modality instead of X-modality (ablation)MUN · RegDBrank-1 (%): 79.95 -> 95.19Improved
add modality prototype scheme in L_ma (ablation)modality alignment loss · SYSU-MM01 all-searchrank-1 (%): 69.02 -> 74.15Improved
add auxiliary bridge scheme in L_ma (ablation)modality alignment loss · SYSU-MM01 all-searchrank-1 (%): 69.02 -> 71.66Improved
add full MUN components to baseline under test-time corruption (ablation)ResNet-50 baseline · SYSU-MM01-Crank-1 (%): 25.92 -> 41.17Improved
use MUN instead of CMT on RegDB infrared to visibleVI-ReID method · RegDB infrared to visiblerank-1 (%): 91.97 (CMT) vs 91.86 (MUN)Hurt

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