Person Re-Identification in Identity Regression Space

TL;DR

This paper introduces an Identity Regression Space (IRS) for person re-identification that offers scalable, efficient, and adaptable solutions suitable for real-world deployment, outperforming existing methods on multiple benchmarks.

Contribution

The paper proposes a novel IRS framework that formulates re-id as a regression problem, enabling high efficiency, incremental learning, and scalability for large populations.

Findings

IRS outperforms state-of-the-art re-id methods on benchmark datasets.

IRS is more scalable and adaptable with rapid model updates.

Active sample selection reduces human labeling effort.

Abstract

Most existing person re-identification (re-id) methods are unsuitable for real-world deployment due to two reasons: Unscalability to large population size, and Inadaptability over time. In this work, we present a unified solution to address both problems. Specifically, we propose to construct an Identity Regression Space (IRS) based on embedding different training person identities (classes) and formulate re-id as a regression problem solved by identity regression in the IRS. The IRS approach is characterised by a closed-form solution with high learning efficiency and an inherent incremental learning capability with human-in-the-loop. Extensive experiments on four benchmarking datasets(VIPeR, CUHK01, CUHK03 and Market-1501) show that the IRS model not only outperforms state-of-the-art re-id methods, but also is more scalable to large re-id population size by rapidly updating model and actively selecting informative samples with reduced human labelling effort.