Dynamic Identity-Guided Attention Network for Visible-Infrared Person Re-identification

TL;DR

This paper introduces DIAN, a novel attention network that enhances cross-modal person re-identification by mining identity-guided, modality-consistent embeddings, achieving state-of-the-art results on SYSU-MM01 and RegDB datasets.

Contribution

The paper proposes a dynamic identity-guided attention network with orthogonal feature fusion and cross embedding balancing loss for improved VI-ReID performance.

Findings

Achieves 86.28% rank-1 accuracy on SYSU-MM01 indoor search.

Outperforms existing methods on SYSU-MM01 and RegDB datasets.

Demonstrates effective bridging of cross-modal discrepancies.

Abstract

Visible-infrared person re-identification (VI-ReID) aims to match people with the same identity between visible and infrared modalities. VI-ReID is a challenging task due to the large differences in individual appearance under different modalities. Existing methods generally try to bridge the cross-modal differences at image or feature level, which lacks exploring the discriminative embeddings. Effectively minimizing these cross-modal discrepancies relies on obtaining representations that are guided by identity and consistent across modalities, while also filtering out representations that are irrelevant to identity. To address these challenges, we introduce a dynamic identity-guided attention network (DIAN) to mine identity-guided and modality-consistent embeddings, facilitating effective bridging the gap between different modalities. Specifically, in DIAN, to pursue a semantically richer representation, we first use orthogonal projection to fuse the features from two connected coarse and fine layers. Furthermore, we first use dynamic convolution kernels to mine identity-guided and modality-consistent representations. More notably, a cross embedding balancing loss is introduced to effectively bridge cross-modal discrepancies by above embeddings. Experimental results on SYSU-MM01 and RegDB datasets show that DIAN achieves state-of-the-art performance. Specifically, for indoor search on SYSU-MM01, our method achieves 86.28% rank-1 accuracy and 87.41% mAP, respectively. Our code will be available soon.