Style Variable and Irrelevant Learning for Generalizable Person Re-identification

TL;DR

This paper introduces SVIL, a method that reduces style bias in person re-identification models by enriching style diversity and combining with meta-learning, significantly improving generalization to unseen domains.

Contribution

We propose a novel Style Jitter Module (SJM) that enhances style diversity and reduces style bias, combined with meta-learning for improved domain generalization in person ReID.

Findings

SVIL outperforms state-of-the-art methods on DG-ReID benchmarks.

The SJM module is plug-and-play and inference cost-free.

Style factors are confirmed as a key part of domain bias in ReID.

Abstract

Recently, due to the poor performance of supervised person re-identification (ReID) to an unseen domain, Domain Generalization (DG) person ReID has attracted a lot of attention which aims to learn a domain-insensitive model and can resist the influence of domain bias. In this paper, we first verify through an experiment that style factors are a vital part of domain bias. Base on this conclusion, we propose a Style Variable and Irrelevant Learning (SVIL) method to eliminate the effect of style factors on the model. Specifically, we design a Style Jitter Module (SJM) in SVIL. The SJM module can enrich the style diversity of the specific source domain and reduce the style differences of various source domains. This leads to the model focusing on identity-relevant information and being insensitive to the style changes. Besides, we organically combine the SJM module with a meta-learning algorithm, maximizing the benefits and further improving the generalization ability of the model. Note that our SJM module is plug-and-play and inference cost-free. Extensive experiments confirm the effectiveness of our SVIL and our method outperforms the state-of-the-art methods on DG-ReID benchmarks by a large margin.