Re-ID done right: towards good practices for person re-identification

TL;DR

This paper establishes effective design and training practices for deep person re-identification models, demonstrating that a carefully crafted simple architecture can outperform complex methods on multiple benchmarks.

Contribution

It provides a comprehensive set of good practices for designing and training deep architectures for person re-identification, emphasizing simplicity and effective training strategies.

Findings

Outperforms state-of-the-art methods on four benchmarks

A simple architecture with good practices achieves high accuracy

Implicit attention mechanism observed in learned representations

Abstract

Training a deep architecture using a ranking loss has become standard for the person re-identification task. Increasingly, these deep architectures include additional components that leverage part detections, attribute predictions, pose estimators and other auxiliary information, in order to more effectively localize and align discriminative image regions. In this paper we adopt a different approach and carefully design each component of a simple deep architecture and, critically, the strategy for training it effectively for person re-identification. We extensively evaluate each design choice, leading to a list of good practices for person re-identification. By following these practices, our approach outperforms the state of the art, including more complex methods with auxiliary components, by large margins on four benchmark datasets. We also provide a qualitative analysis of our trained representation which indicates that, while compact, it is able to capture information from localized and discriminative regions, in a manner akin to an implicit attention mechanism.