Discovering Underlying Person Structure Pattern with Relative Local Distance for Person Re-identification

TL;DR

This paper introduces a novel Relative Local Distance (RLD) method that enhances person re-identification by guiding CNNs to learn structure-aware features without extra annotations, improving accuracy and training speed.

Contribution

The paper proposes the first use of a relative local distance constraint within CNNs to incorporate person structure information for re-ID, boosting feature representation and training efficiency.

Findings

RLD improves re-ID accuracy on multiple datasets.

It accelerates deep network training compared to traditional methods.

The method effectively captures underlying person structure without extra annotations.

Abstract

Modeling the underlying person structure for person re-identification (re-ID) is difficult due to diverse deformable poses, changeable camera views and imperfect person detectors. How to exploit underlying person structure information without extra annotations to improve the performance of person re-ID remains largely unexplored. To address this problem, we propose a novel Relative Local Distance (RLD) method that integrates a relative local distance constraint into convolutional neural networks (CNNs) in an end-to-end way. It is the first time that the relative local constraint is proposed to guide the global feature representation learning. Specially, a relative local distance matrix is computed by using feature maps and then regarded as a regularizer to guide CNNs to learn a structure-aware feature representation. With the discovered underlying person structure, the RLD method builds a bridge between the global and local feature representation and thus improves the capacity of feature representation for person re-ID. Furthermore, RLD also significantly accelerates deep network training compared with conventional methods. The experimental results show the effectiveness of RLD on the CUHK03, Market-1501, and DukeMTMC-reID datasets. Code is available at \url{https://github.com/Wanggcong/RLD_codes}.