Domain Generalized Person Re-Identification via Cross-Domain Episodic Learning

TL;DR

This paper introduces a novel episodic learning approach for domain-generalized person re-identification, enabling recognition across unseen camera domains without target domain training data, outperforming existing methods.

Contribution

It proposes a cross-domain episodic learning scheme that enhances meta-learning to learn domain-invariant features without target domain data.

Findings

Outperforms state-of-the-art on four benchmark datasets.

Effectively learns domain-invariant features without target data.

Demonstrates robustness across diverse unseen domains.

Abstract

Aiming at recognizing images of the same person across distinct camera views, person re-identification (re-ID) has been among active research topics in computer vision. Most existing re-ID works require collection of a large amount of labeled image data from the scenes of interest. When the data to be recognized are different from the source-domain training ones, a number of domain adaptation approaches have been proposed. Nevertheless, one still needs to collect labeled or unlabelled target-domain data during training. In this paper, we tackle an even more challenging and practical setting, domain generalized (DG) person re-ID. That is, while a number of labeled source-domain datasets are available, we do not have access to any target-domain training data. In order to learn domain-invariant features without knowing the target domain of interest, we present an episodic learning scheme which advances meta learning strategies to exploit the observed source-domain labeled data. The learned features would exhibit sufficient domain-invariant properties while not overfitting the source-domain data or ID labels. Our experiments on four benchmark datasets confirm the superiority of our method over the state-of-the-arts.