Making Person Search Enjoy the Merits of Person Re-identification

TL;DR

This paper introduces TDN, a novel framework that leverages advanced Re-ID models to enhance one-step person search by disentangling subtasks, transferring knowledge, and exploiting context for improved accuracy.

Contribution

The paper proposes TDN, a framework that significantly boosts person search performance by integrating Re-ID knowledge transfer and context exploitation in a unified model.

Findings

TDN outperforms existing methods on public datasets.

Knowledge transfer from Re-ID models improves search accuracy.

Context-based ranking enhances retrieval results.

Abstract

Person search is an extended task of person re-identification (Re-ID). However, most existing one-step person search works have not studied how to employ existing advanced Re-ID models to boost the one-step person search performance due to the integration of person detection and Re-ID. To address this issue, we propose a faster and stronger one-step person search framework, the Teacher-guided Disentangling Networks (TDN), to make the one-step person search enjoy the merits of the existing Re-ID researches. The proposed TDN can significantly boost the person search performance by transferring the advanced person Re-ID knowledge to the person search model. In the proposed TDN, for better knowledge transfer from the Re-ID teacher model to the one-step person search model, we design a strong one-step person search base framework by partially disentangling the two subtasks. Besides, we propose a Knowledge Transfer Bridge module to bridge the scale gap caused by different input formats between the Re-ID model and one-step person search model. During testing, we further propose the Ranking with Context Persons strategy to exploit the context information in panoramic images for better retrieval. Experiments on two public person search datasets demonstrate the favorable performance of the proposed method.