DDRN:a Data Distribution Reconstruction Network for Occluded Person Re-Identification

TL;DR

This paper introduces DDRN, a generative network that reconstructs data distribution to improve occluded person re-identification, effectively filtering irrelevant information and handling complex feature spaces, resulting in state-of-the-art performance.

Contribution

The paper proposes DDRN, a novel data distribution reconstruction network, and HS-Arcface loss for better feature space approximation in occluded person ReID.

Findings

Achieved 62.4% mAP on Occluded-Duke dataset, surpassing previous methods.

Improved rank-1 accuracy to 71.3%, demonstrating enhanced identification performance.

Effectively filters irrelevant background and occlusion information in ReID tasks.

Abstract

In occluded person re-identification(ReID), severe occlusions lead to a significant amount of irrelevant information that hinders the accurate identification of individuals. These irrelevant cues primarily stem from background interference and occluding interference, adversely affecting the final retrieval results. Traditional discriminative models, which rely on the specific content and positions of the images, often misclassify in cases of occlusion. To address these limitations, we propose the Data Distribution Reconstruction Network (DDRN), a generative model that leverages data distribution to filter out irrelevant details, enhancing overall feature perception ability and reducing irrelevant feature interference. Additionally, severe occlusions lead to the complexity of the feature space. To effectively handle this, we design a multi-center approach through the proposed Hierarchical SubcenterArcface (HS-Arcface) loss function, which can better approximate complex feature spaces. On the Occluded-Duke dataset, we achieved a mAP of 62.4\% (+1.1\%) and a rank-1 accuracy of 71.3\% (+0.6\%), surpassing the latest state-of-the-art methods(FRT) significantly.