Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro

TL;DR

This paper introduces a semi-supervised approach using GAN-generated unlabeled samples with label smoothing regularization to enhance person re-identification performance without extra data collection.

Contribution

It proposes LSRO, a novel regularization method for utilizing GAN-generated unlabeled data to improve CNN-based person re-ID models.

Findings

Improved rank-1 accuracy on three large datasets: Market-1501, CUHK03, DukeMTMC-reID.

GAN data augmentation enhances discriminative ability of CNN embeddings.

Method also benefits fine-grained bird recognition with a +0.6% accuracy gain.

Abstract

The main contribution of this paper is a simple semi-supervised pipeline that only uses the original training set without collecting extra data. It is challenging in 1) how to obtain more training data only from the training set and 2) how to use the newly generated data. In this work, the generative adversarial network (GAN) is used to generate unlabeled samples. We propose the label smoothing regularization for outliers (LSRO). This method assigns a uniform label distribution to the unlabeled images, which regularizes the supervised model and improves the baseline. We verify the proposed method on a practical problem: person re-identification (re-ID). This task aims to retrieve a query person from other cameras. We adopt the deep convolutional generative adversarial network (DCGAN) for sample generation, and a baseline convolutional neural network (CNN) for representation learning. Experiments show that adding the GAN-generated data effectively improves the discriminative ability of learned CNN embeddings. On three large-scale datasets, Market-1501, CUHK03 and DukeMTMC-reID, we obtain +4.37%, +1.6% and +2.46% improvement in rank-1 precision over the baseline CNN, respectively. We additionally apply the proposed method to fine-grained bird recognition and achieve a +0.6% improvement over a strong baseline. The code is available at https://github.com/layumi/Person-reID_GAN.