DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations

TL;DR

This paper introduces DomainMix, a scalable, annotation-free framework that leverages synthetic and unlabeled real-world data to train highly generalizable person re-identification models, reducing the need for costly human annotations.

Contribution

The paper proposes a novel domain-mixing framework with domain-invariant feature learning and a domain balance loss to improve generalization without human labels.

Findings

Achieves comparable performance to fully supervised models.

Sets new state-of-the-art in cross-dataset person re-identification.

Effectively reduces domain gap between synthetic and real data.

Abstract

Existing person re-identification models often have low generalizability, which is mostly due to limited availability of large-scale labeled data in training. However, labeling large-scale training data is very expensive and time-consuming, while large-scale synthetic dataset shows promising value in learning generalizable person re-identification models. Therefore, in this paper a novel and practical person re-identification task is proposed,i.e. how to use labeled synthetic dataset and unlabeled real-world dataset to train a universal model. In this way, human annotations are no longer required, and it is scalable to large and diverse real-world datasets. To address the task, we introduce a framework with high generalizability, namely DomainMix. Specifically, the proposed method firstly clusters the unlabeled real-world images and selects the reliable clusters. During training, to address the large domain gap between two domains, a domain-invariant feature learning method is proposed, which introduces a new loss,i.e. domain balance loss, to conduct an adversarial learning between domain-invariant feature learning and domain discrimination, and meanwhile learns a discriminative feature for person re-identification. This way, the domain gap between synthetic and real-world data is much reduced, and the learned feature is generalizable thanks to the large-scale and diverse training data. Experimental results show that the proposed annotation-free method is more or less comparable to the counterpart trained with full human annotations, which is quite promising. In addition, it achieves the current state of the art on several person re-identification datasets under direct cross-dataset evaluation.