Re-identify people in new domains with unsupervised learning by rewinding video with CycAs
Re-identify people in new domains with unsupervised learning by rewinding video with CycAs
Generalizable Re-Identification from Videos with Cycle Association
arXiv paper abstract https://arxiv.org/abs/2211.03663
arXiv PDF paper https://arxiv.org/pdf/2211.03663.pdf
… interested in learning a generalizable person re-identification (re-ID) representation from unlabeled videos.
… aim to learn a representation in an unsupervised manner and directly use the learned representation for re-ID in novel domains.
… First … propose Cycle Association (CycAs), a scalable self-supervised learning method for re-ID with low training complexity … second … construct a large-scale unlabeled re-ID dataset named LMP-video
… CycAs learns re-ID features by enforcing cycle consistency of instance association between temporally successive video frame pairs, and the training cost is merely linear to the data size, making large-scale training possible.
… Trained on LMP-video, … show that CycAs learns good generalization towards novel domains.
… sometimes even outperform supervised domain generalizable models … CycAs … surpassing state-of-the-art supervised DG re-ID methods …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
