Learning a Discriminative Null Space for Person Re-identification

TL;DR

This paper introduces a null space approach for person re-identification that effectively addresses the small sample size problem, leading to superior performance across multiple benchmarks.

Contribution

It proposes a discriminative null space method for re-id that is simple, computationally efficient, and outperforms existing state-of-the-art techniques.

Findings

Outperforms state-of-the-art methods on five benchmarks.

Efficient closed-form solution with fixed dimension.

Effectively handles small sample size problem.

Abstract

Most existing person re-identification (re-id) methods focus on learning the optimal distance metrics across camera views. Typically a person's appearance is represented using features of thousands of dimensions, whilst only hundreds of training samples are available due to the difficulties in collecting matched training images. With the number of training samples much smaller than the feature dimension, the existing methods thus face the classic small sample size (SSS) problem and have to resort to dimensionality reduction techniques and/or matrix regularisation, which lead to loss of discriminative power. In this work, we propose to overcome the SSS problem in re-id distance metric learning by matching people in a discriminative null space of the training data. In this null space, images of the same person are collapsed into a single point thus minimising the within-class scatter to the extreme and maximising the relative between-class separation simultaneously. Importantly, it has a fixed dimension, a closed-form solution and is very efficient to compute. Extensive experiments carried out on five person re-identification benchmarks including VIPeR, PRID2011, CUHK01, CUHK03 and Market1501 show that such a simple approach beats the state-of-the-art alternatives, often by a big margin.