Learning Commonality, Divergence and Variety for Unsupervised Visible-Infrared Person Re-identification

TL;DR

This paper introduces a novel unsupervised visible-infrared person re-identification method that leverages progressive contrastive learning with hard and dynamic prototypes to better capture commonality, divergence, and variety among images.

Contribution

It proposes a progressive contrastive learning framework that explicitly models divergence and variety using hard and dynamic prototypes, improving USVI-ReID performance.

Findings

Outperforms existing methods on SYSU-MM01 and RegDB datasets.

Effectively emphasizes divergence and variety in feature representations.

Demonstrates robustness and improved accuracy in unsupervised cross-modality matching.

Abstract

Unsupervised visible-infrared person re-identification (USVI-ReID) aims to match specified people in infrared images to visible images without annotations, and vice versa. USVI-ReID is a challenging yet under-explored task. Most existing methods address the USVI-ReID using cluster-based contrastive learning, which simply employs the cluster center as a representation of a person. However, the cluster center primarily focuses on commonality, overlooking divergence and variety. To address the problem, we propose a Progressive Contrastive Learning with Hard and Dynamic Prototypes method for USVI-ReID. In brief, we generate the hard prototype by selecting the sample with the maximum distance from the cluster center. We theoretically show that the hard prototype is used in the contrastive loss to emphasize divergence. Additionally, instead of rigidly aligning query images to a specific prototype, we generate the dynamic prototype by randomly picking samples within a cluster. The dynamic prototype is used to encourage the variety. Finally, we introduce a progressive learning strategy to gradually shift the model's attention towards divergence and variety, avoiding cluster deterioration. Extensive experiments conducted on the publicly available SYSU-MM01 and RegDB datasets validate the effectiveness of the proposed method.