Extended Cross-Modality United Learning for Unsupervised Visible-Infrared Person Re-identification

TL;DR

This paper introduces ECUL, a novel framework for unsupervised visible-infrared person re-identification that effectively learns modality-invariant features by integrating advanced clustering and memory strategies, outperforming some supervised methods.

Contribution

The paper proposes ECUL, a new framework combining intra- and inter-modality clustering with a two-step memory update, enhancing unsupervised cross-modality feature learning.

Findings

ECUL outperforms existing unsupervised methods on SYSU-MM01 and RegDB datasets.

ECUL achieves competitive results compared to some supervised approaches.

The framework effectively reduces inter-modality gap and noise in clustering.

Abstract

Unsupervised learning visible-infrared person re-identification (USL-VI-ReID) aims to learn modality-invariant features from unlabeled cross-modality datasets and reduce the inter-modality gap. However, the existing methods lack cross-modality clustering or excessively pursue cluster-level association, which makes it difficult to perform reliable modality-invariant features learning. To deal with this issue, we propose a Extended Cross-Modality United Learning (ECUL) framework, incorporating Extended Modality-Camera Clustering (EMCC) and Two-Step Memory Updating Strategy (TSMem) modules. Specifically, we design ECUL to naturally integrates intra-modality clustering, inter-modality clustering and inter-modality instance selection, establishing compact and accurate cross-modality associations while reducing the introduction of noisy labels. Moreover, EMCC captures and filters the neighborhood relationships by extending the encoding vector, which further promotes the learning of modality-invariant and camera-invariant knowledge in terms of clustering algorithm. Finally, TSMem provides accurate and generalized proxy points for contrastive learning by updating the memory in stages. Extensive experiments results on SYSU-MM01 and RegDB datasets demonstrate that the proposed ECUL shows promising performance and even outperforms certain supervised methods.