Sampling Agnostic Feature Representation for Long-Term Person Re-identification

TL;DR

This paper introduces SirNet, a novel person re-identification framework that learns disentangled features independently of sample sampling, improving discriminability and robustness across diverse datasets.

Contribution

The paper proposes a sampling independent feature learning method with a maximum discrepancy loss, enhancing re-identification performance beyond existing models.

Findings

Outperforms prior state-of-the-art on benchmark datasets

Learns disentangled features independent of sample characteristics

Generates hard negatives and positives for improved discriminability

Abstract

Person re-identification is a problem of identifying individuals across non-overlapping cameras. Although remarkable progress has been made in the re-identification problem, it is still a challenging problem due to appearance variations of the same person as well as other people of similar appearance. Some prior works solved the issues by separating features of positive samples from features of negative ones. However, the performances of existing models considerably depend on the characteristics and statistics of the samples used for training. Thus, we propose a novel framework named sampling independent robust feature representation network (SirNet) that learns disentangled feature embedding from randomly chosen samples. A carefully designed sampling independent maximum discrepancy loss is introduced to model samples of the same person as a cluster. As a result, the proposed framework can generate additional hard negatives/positives using the learned features, which results in better discriminability from other identities. Extensive experimental results on large-scale benchmark datasets verify that the proposed model is more effective than prior state-of-the-art models.