AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Camera-Driven Representation Learning for Unsupervised Domain Adaptive Person Re-identification

Geon Lee, Sanghoon Lee, Dohyung Kim, Younghoon Shin, Yongsang Yoon, Bumsub Ham

ICCV 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Geometric alignment and synthetic-to-real adaptation · 59 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 camera-driven curriculum (CaCL) over cross-entropy+triplet loss (ablation)UDA reID with ResNet-50 · Market1501-to-MSMT17mAP: 23.4 -> 31.1Improved
use random curriculum sequence instead of none (ablation)CTL baseline · Market1501-to-MSMT17mAP: 23.4 -> 24.4Improved
replace random curriculum with MMD-based camera-driven scheduler (with CD loss) (ablation)CaCL · Market1501-to-MSMT17mAP: 30.4 -> 36.5Improved
add camera-diversity loss on top of CaCL (ablation)CaCL · Market1501-to-MSMT17mAP: 31.1 -> 36.5Improved
add camera-diversity loss without curriculum (ablation)baseline CTL · MSMT17-to-Market1501mAP: 79.2 -> 81.8Improved
CD loss instead of UGID loss (ablation)UDA reID · Market1501-to-MSMT17mAP: 32.3 -> 36.5Improved
combine CD loss with UGID weights (ablation)CaCL with CD loss · Market1501-to-MSMT17mAP: 36.5 -> 38.4Improved
combine CD loss with UGID weights (ablation)CaCL with CD loss · Unreal-to-MSMT17mAP: 40.3 -> 41.2Improved

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