3C: Confidence-Guided Clustering and Contrastive Learning for Unsupervised Person Re-Identification

TL;DR

This paper addresses unsupervised person Re-ID by mitigating pseudo-label noise and camera bias through a confidence-guided framework named 3C. It combines three mechanisms—Harmonic Discrepancy Clustering (HDC) for robust clustering, Camera Information Entropy (CIE) to weight cross-camera diversity during forward learning, and Confidence Integrated Harmonic Discrepancy (CHD) to guide memory updates with camera-aware hard samples. Empirical results on Market-1501, MSMT17, and VeRi-776 demonstrate state-of-the-art or competitive performance, especially in complex scenarios like MSMT17 and VeRi-776, with ablations validating each component’s contribution. The approach offers practical impact for unsupervised Re-ID by enhancing label quality, learning efficiency, and robustness to cross-camera variability, and suggests future work on broader USL tasks and backbone optimization.

Abstract

Unsupervised person re-identification (Re-ID) aims to learn a feature network with cross-camera retrieval capability in unlabelled datasets. Although the pseudo-label based methods have achieved great progress in Re-ID, their performance in the complex scenario still needs to sharpen up. In order to reduce potential misguidance, including feature bias, noise pseudo-labels and invalid hard samples, accumulated during the learning process, in this pa per, a confidence-guided clustering and contrastive learning (3C) framework is proposed for unsupervised person Re-ID. This 3C framework presents three confidence degrees. i) In the clustering stage, the confidence of the discrepancy between samples and clusters is proposed to implement a harmonic discrepancy clustering algorithm (HDC). ii) In the forward-propagation training stage, the confidence of the camera diversity of a cluster is evaluated via a novel camera information entropy (CIE). Then, the clusters with high CIE values will play leading roles in training the model. iii) In the back-propagation training stage, the confidence of the hard sample in each cluster is designed and further used in a confidence integrated harmonic discrepancy (CHD), to select the informative sample for updating the memory in contrastive learning. Extensive experiments on three popular Re-ID benchmarks demonstrate the superiority of the proposed framework. Particularly, the 3C framework achieves state-of-the-art results: 86.7%/94.7%, 45.3%/73.1% and 47.1%/90.6% in terms of mAP/Rank-1 accuracy on Market-1501, the com plex datasets MSMT17 and VeRi-776, respectively. Code is available at https://github.com/stone5265/3C-reid.

Figures (7)

• Figure 1: Comparison among $K$ -means, DBSCAN and HDC.
• Figure 2: An overview on the proposed 3C framework.
• Figure 3: Illustration of CIE for real IDs.
• Figure 4: Exemplar comparison among different memory hard sampling update strategies. Hollow points denote samples. Black solid points denote the cluster centroids in memory. Green and orange solid points denote the two camera centres, repectively. Black solid lines denote the equidistant lines. Black dashed arrows denote the directions of memory update.
• Figure 5: Evaluation on different number of clusters.