Pseudo Label Rectification With Joint Camera Shift Adaptation and Outlier Progressive Recycling for Unsupervised Person Re-Identification

Person re-identification (re-ID) has many applications in intelligent transportation systems. Clustering-based methods, which alternate between the generation of pseudo labels via clustering and the optimization of the feature extractor, have obtained leading performance in unsupervised person re-ID. But there are still two issues not well addressed: 1) Most methods measure the feature similarity without considering the domain shift between cameras, degrading the clustering performance. 2) Outliers, which usually correspond to hard samples with large discrepancy from other images of the identical person, are in most cases directly excluded from the network training. To tackle the above issues, this paper proposes a plug-and-play pseudo label rectification framework, which jointly utilizes CAmera Shift adapTation module and Outlier progressive Recycling strategy ( $CASTOR$ ) to improve the quality of pseudo labels from both pre-clustering and post-clustering. Specifically, we first compute the camera similarity of two samples by utilizing a pretrained camera classification network and subtract the feature similarity by the camera similarity, the value of which is weighted in an exponential decay manner throughout the network training, in order to adaptively remedy the adverse impact of inter-camera distribution shift upon clustering. Besides, we carefully design an outlier progressive recycling strategy to reassign part of the outliers into the clustered groups to make full use of the useful information of outliers. Extensive experiments on three large scale unsupervised and unsupervised domain adaptive (UDA) person re-ID benchmarks validate the effectiveness of $CASTOR$ and its wide compatibility with the state-of-the-art clustering-based methods.

I. Introduction

Person re-identification (re-ID) aims to associate the images of the identical person captured by different cameras. Due to the urgent demand of public safety and increasing number of surveillance cameras in university campuses, theme parks, transportation roads, etc, the development of this technology is conducive to intelligent transportation [4], [5] and video surveillance system design [6], [7]. It is a quite challenging task due to the complex intra-class and inter-class variations caused by illumination change, camera viewpoint shift, human pose variation, occlusion, etc. With the incredible power of deep neural networks, the performance of fully supervised person re-ID [8], [9], [10], [11] on public benchmarks [2], [12], [13] has made significant progress. However, in practical application scenarios, supervised person re-ID suffers from the lack of sufficient labelled training data due to the high cost of manual annotation. Recently, researchers pay more attention to the unsupervised person re-ID and the unsupervised domain adaptive (UDA) person re-ID [3], [14], [15], [16], [17], [18], [19], [20], [21]. The former task only utilizes the unlabeled data collected in the target scene to train the re-ID model. The latter task jointly utilizes the unlabeled data in the target domain and labelled data from the source domain to improve the person re-ID performance.