Collaborative Contrastive Refining for Weakly Supervised Person Search

Authors

Chengyou Jia, Xi'an Jiaotong University
Minnan Luo, Xi'an Jiaotong University
Caixia Yan, Xi'an Jiaotong University
Linchao Zhu, College of Computer Science and Technology, Zhejiang University
Xiaojun Chang, University of Technology Sydney & Mohamed bin Zayed University of Artificial IntelligenceFollow
Qinghua Zheng, Xi'an Jiaotong University

Document Type

Article

Publication Title

IEEE Transactions on Image Processing

Abstract

Weakly supervised person search involves training a model with only bounding box annotations, without human-annotated identities. Clustering algorithms are commonly used to assign pseudo-labels to facilitate this task. However, inaccurate pseudo-labels and imbalanced identity distributions can result in severe label and sample noise. In this work, we propose a novel Collaborative Contrastive Refining (CCR) weakly-supervised framework for person search that jointly refines pseudo-labels and the sample-learning process with different contrastive strategies. Specifically, we adopt a hybrid contrastive strategy that leverages both visual and context clues to refine pseudo-labels, and leverage the sample-mining and noise-contrastive strategy to reduce the negative impact of imbalanced distributions by distinguishing positive samples and noise samples. Our method brings two main advantages: 1) it facilitates better clustering results for refining pseudo-labels by exploring the hybrid similarity; 2) it is better at distinguishing query samples and noise samples for refining the sample-learning process. Extensive experiments demonstrate the superiority of our approach over the state-of-the-art weakly supervised methods by a large margin (more than 3%mAP on CUHK-SYSU). Moreover, by leveraging more diverse unlabeled data, our method achieves comparable or even better performance than the state-of-the-art supervised methods.

First Page

4951

Last Page

4963

DOI

10.1109/TIP.2023.3308393

Publication Date

8-29-2023

Keywords

clustering algorithm, Person search, unsupervised person Re-ID, weakly supervised learning

Comments

IR Deposit conditions:

OA version (pathway a) Accepted version

No embargo

When accepted for publication, set statement to accompany deposit (see policy)

Must link to publisher version with DOI

Publisher copyright and source must be acknowledged

Recommended Citation

C. Jia, M. Luo, C. Yan, L. Zhu, X. Chang and Q. Zheng, "Collaborative Contrastive Refining for Weakly Supervised Person Search," in IEEE Transactions on Image Processing, vol. 32, pp. 4951-4963, 2023, doi: 10.1109/TIP.2023.3308393.

Additional Links

https://doi.org/10.1109/TIP.2023.3308393