AI Paper Atlas

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

Scale-Aware Domain Harmonization for Domain Adaptation Person Search

Huibing Wang, Guojian Zhao, Jinjia Peng, Linfeng Qi, Mingze Yao, Jiqing Zhang

ICML 2026 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Feature alignment for unsupervised adaptation

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 SIA, SCR and BCR to DSCA baseline (ablation)SCALE · PRWmAP: 39.9 -> 41.7Improved
add SIA, SCR and BCR to DSCA baseline (ablation)SCALE · CUHK-SYSUmAP: 80.2 -> 82.3Improved
add SIA module (ablation)SCALE · PRWmAP: 41.1 -> 41.7Improved
use 2 sub-clusters instead of 1 in BCR split (ablation)BCR · CUHK-SYSUmAP: 80.4 -> 81.2Improved
use 3 sub-clusters instead of 2 in BCR split (ablation)BCR · CUHK-SYSUmAP: 81.2 -> 79.8Hurt
add both split and merge branches in BCR (ablation)SCALE · PRWmAP: 39.9 -> 40.8Improved
replace SDH with FPN-style modification (ablation)SCALE · CUHK-SYSUmAP: 82.3 -> 81.1Hurt
replace SDH with multi-scale training/testing (ablation)SCALE · CUHK-SYSUmAP: 82.3 -> 80.6Hurt
use both spatial and semantic paths instead of double spatial (ablation)SDH · CUHK-SYSUmAP: 81.0 -> 82.3Improved
evaluate on scale-mismatched subset (area < 10k)SCALE vs baseline · PRW-style scale-mismatched subsetmAP: 53.53 -> 57.49Improved
add SDH moduleSCALE · computational costparameters / GMac: 56.19M -> 58.64M; 414.09 -> 459.90 GMacHurt

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