Generalizable person Re-Identification (ReID) has attracted growing attention in recent computer vision community. In this work, we construct a structural causal model among identity labels, identity-specific factors (clothes/shoes color etc), and domain-specific factors (background, viewpoints etc). According to the causal analysis, we propose a novel Domain Invariant Representation Learning for generalizable person Re-Identification (DIR-ReID) framework. Specifically, we first propose to disentangle the identity-specific and domain-specific feature spaces, based on which we propose an effective algorithmic implementation for backdoor adjustment, essentially serving as a causal intervention towards the SCM. Extensive experiments have been conducted, showing that DIR-ReID outperforms state-of-the-art methods on large-scale domain generalization ReID benchmarks.
@article{arxiv.2103.15890,
title = {Learning Domain Invariant Representations for Generalizable Person Re-Identification},
author = {Yi-Fan Zhang and Zhang Zhang and Da Li and Zhen Jia and Liang Wang and Tieniu Tan},
journal= {arXiv preprint arXiv:2103.15890},
year = {2022}
}