Title:From Synthetic to Real: Unveiling the Power of Synthetic Data for Video Person Re-ID

Abstract:In this study, we investigate the novel challenge of cross-domain video-based person re-identification (Re-ID). Here, we utilize synthetic video datasets as the source domain for training and real-world videos for testing, notably reducing the reliance on expensive real data acquisition and annotation. To harness the potential of synthetic data, we first propose a self-supervised domain-invariant feature learning strategy for both static and dynamic (temporal) features. Additionally, to enhance person identification accuracy in the target domain, we propose a mean-teacher scheme incorporating a self-supervised ID consistency loss. Experimental results across five real datasets validate the rationale behind cross-synthetic-real domain adaptation and demonstrate the efficacy of our method. Notably, the discovery that synthetic data outperforms real data in the cross-domain scenario is a surprising outcome. The code and data will be publicly available at this https URL.