Identity-aware Feature Decoupling Learning for Clothing-change Person Re-identification

TL;DR

This paper introduces an identity-aware feature decoupling framework for clothing-change person re-identification, effectively extracting identity-related features despite clothing variations using a dual stream architecture with attention mechanisms.

Contribution

The proposed framework employs a dual stream architecture with clothing-masked inputs and a clothing bias diminishing module to better extract identity features in CC Re-ID tasks.

Findings

Outperforms baseline models on multiple CC Re-ID datasets

Effectively highlights identity-related regions despite clothing changes

Reduces semantic gap between clothing-relevant and identity features

Abstract

Clothing-change person re-identification (CC Re-ID) has attracted increasing attention in recent years due to its application prospect. Most existing works struggle to adequately extract the ID-related information from the original RGB images. In this paper, we propose an Identity-aware Feature Decoupling (IFD) learning framework to mine identity-related features. Particularly, IFD exploits a dual stream architecture that consists of a main stream and an attention stream. The attention stream takes the clothing-masked images as inputs and derives the identity attention weights for effectively transferring the spatial knowledge to the main stream and highlighting the regions with abundant identity-related information. To eliminate the semantic gap between the inputs of two streams, we propose a clothing bias diminishing module specific to the main stream to regularize the features of clothing-relevant regions. Extensive experimental results demonstrate that our framework outperforms other baseline models on several widely-used CC Re-ID datasets.