Fully Unsupervised Person Re-identification viaSelective Contrastive Learning

TL;DR

This paper introduces a novel unsupervised person re-identification method using selective contrastive learning with multiple positives, adaptive negatives, and combined global-local features, achieving superior results.

Contribution

It proposes a new selective contrastive learning framework with dynamic dictionaries and combined global-local features for improved unsupervised person ReID.

Findings

Outperforms state-of-the-art unsupervised ReID methods.

Effective use of multiple positives and adaptive negatives.

Utilizes global and local features for better discrimination.

Abstract

Person re-identification (ReID) aims at searching the same identity person among images captured by various cameras. Unsupervised person ReID attracts a lot of attention recently, due to it works without intensive manual annotation and thus shows great potential of adapting to new conditions. Representation learning plays a critical role in unsupervised person ReID. In this work, we propose a novel selective contrastive learning framework for unsupervised feature learning. Specifically, different from traditional contrastive learning strategies, we propose to use multiple positives and adaptively sampled negatives for defining the contrastive loss, enabling to learn a feature embedding model with stronger identity discriminative representation. Moreover, we propose to jointly leverage global and local features to construct three dynamic dictionaries, among which the global and local memory banks are used for pairwise similarity computation and the mixture memory bank are used for contrastive loss definition. Experimental results demonstrate the superiority of our method in unsupervised person ReID compared with the state-of-the-arts.