Deep High-Resolution Representation Learning for Cross-Resolution Person Re-identification

TL;DR

This paper introduces a novel deep learning framework, PS-HRNet, designed to improve cross-resolution person re-identification by restoring low-resolution images and extracting discriminative features, achieving state-of-the-art results on multiple datasets.

Contribution

The paper proposes a new pseudo-siamese high-resolution network with a VDSR-CA module for better resolution restoration and feature extraction in cross-resolution person re-ID.

Findings

Improves Rank-1 accuracy by up to 6.2% on benchmark datasets.

Effectively reduces feature distribution differences between low- and high-resolution images.

Outperforms existing state-of-the-art methods in cross-resolution person re-ID.

Abstract

Person re-identification (re-ID) tackles the problem of matching person images with the same identity from different cameras. In practical applications, due to the differences in camera performance and distance between cameras and persons of interest, captured person images usually have various resolutions. We name this problem as Cross-Resolution Person Re-identification which brings a great challenge for matching correctly. In this paper, we propose a Deep High-Resolution Pseudo-Siamese Framework (PS-HRNet) to solve the above problem. Specifically, in order to restore the resolution of low-resolution images and make reasonable use of different channel information of feature maps, we introduce and innovate VDSR module with channel attention (CA) mechanism, named as VDSR-CA. Then we reform the HRNet by designing a novel representation head to extract discriminating features, named as HRNet-ReID. In addition, a pseudo-siamese framework is constructed to reduce the difference of feature distributions between low-resolution images and high-resolution images. The experimental results on five cross-resolution person datasets verify the effectiveness of our proposed approach. Compared with the state-of-the-art methods, our proposed PS-HRNet improves 3.4\%, 6.2\%, 2.5\%,1.1\% and 4.2\% at Rank-1 on MLR-Market-1501, MLR-CUHK03, MLR-VIPeR, MLR-DukeMTMC-reID, and CAVIAR datasets, respectively. Our code is available at \url{https://github.com/zhguoqing}.