Unsupervised Domain-Adaptive Person Re-identification Based on Attributes

TL;DR

This paper introduces an unsupervised domain adaptation framework for person re-identification that leverages pedestrian attribute annotations from different datasets to improve ReID performance across domains.

Contribution

It proposes an adversarial discriminative domain adaptation method to transfer attribute-related features, enabling domain-invariant semantic attribute encoding for ReID.

Findings

Effective cross-domain ReID achieved on large-scale datasets.

Attribute transfer improves ReID accuracy across different datasets.

The framework outperforms existing unsupervised methods.

Abstract

Pedestrian attributes, e.g., hair length, clothes type and color, locally describe the semantic appearance of a person. Training person re-identification (ReID) algorithms under the supervision of such attributes have proven to be effective in extracting local features which are important for ReID. Unlike person identity, attributes are consistent across different domains (or datasets). However, most of ReID datasets lack attribute annotations. On the other hand, there are several datasets labeled with sufficient attributes for the case of pedestrian attribute recognition. Exploiting such data for ReID purpose can be a way to alleviate the shortage of attribute annotations in ReID case. In this work, an unsupervised domain adaptive ReID feature learning framework is proposed to make full use of attribute annotations. We propose to transfer attribute-related features from their original domain to the ReID one: to this end, we introduce an adversarial discriminative domain adaptation method in order to learn domain invariant features for encoding semantic attributes. Experiments on three large-scale datasets validate the effectiveness of the proposed ReID framework.