Domain Generalizable Person Search Using Unreal Dataset

arXiv:2404.00626v1 Announce Type: new
Abstract: Collecting and labeling real datasets to train the person search networks not only requires a lot of time and effort, but also accompanies privacy issues. The weakly-supervised and unsupervised domain adaptation methods have been proposed to alleviate the labeling burden for target datasets, however, their generalization capability is limited. We introduce a novel person search method based on the domain generalization framework, that uses an automatically labeled unreal dataset only for training but is applicable …

abstract arxiv capability cs.cv dataset datasets domain domain adaptation however labeling networks person privacy search train type unreal unsupervised weakly-supervised