Person Re-Identification by Deep Joint Learning of Multi-Loss Classification

TL;DR

This paper introduces a novel CNN architecture called JLML that jointly learns local and global features for person re-identification, improving accuracy by leveraging their complementary effects using multi-loss optimization.

Contribution

The paper proposes a new CNN model for joint local and global feature learning with multi-loss optimization, enhancing person re-id performance over existing methods.

Findings

JLML outperforms state-of-the-art re-id methods on five benchmarks.

Joint learning of local and global features improves discriminative power.

Multi-loss training effectively captures complementary features for re-id.

Abstract

Existing person re-identification (re-id) methods rely mostly on either localised or global feature representation alone. This ignores their joint benefit and mutual complementary effects. In this work, we show the advantages of jointly learning local and global features in a Convolutional Neural Network (CNN) by aiming to discover correlated local and global features in different context. Specifically, we formulate a method for joint learning of local and global feature selection losses designed to optimise person re-id when using only generic matching metrics such as the L2 distance. We design a novel CNN architecture for Jointly Learning Multi-Loss (JLML) of local and global discriminative feature optimisation subject concurrently to the same re-id labelled information. Extensive comparative evaluations demonstrate the advantages of this new JLML model for person re-id over a wide range of state-of-the-art re-id methods on five benchmarks (VIPeR, GRID, CUHK01, CUHK03, Market-1501).