Investigating the Potential of Auxiliary-Classifier GANs for Image Classification in Low Data Regimes

TL;DR

This paper explores the use of Auxiliary-Classifier GANs as an integrated approach for image classification, especially effective in low data scenarios, by modifying training schemes and loss functions to improve stability and performance.

Contribution

It introduces modifications to AC-GANs, including new sampling schemes and Wasserstein loss, demonstrating their effectiveness in low data image classification tasks.

Findings

AC-GANs achieve competitive accuracy with CNNs in low data regimes.

Modified AC-GANs show improved training stability and performance.

The approach is effective across various image resolutions and complexities.

Abstract

Generative Adversarial Networks (GANs) have shown promise in augmenting datasets and boosting convolutional neural networks' (CNN) performance on image classification tasks. But they introduce more hyperparameters to tune as well as the need for additional time and computational power to train supplementary to the CNN. In this work, we examine the potential for Auxiliary-Classifier GANs (AC-GANs) as a 'one-stop-shop' architecture for image classification, particularly in low data regimes. Additionally, we explore modifications to the typical AC-GAN framework, changing the generator's latent space sampling scheme and employing a Wasserstein loss with gradient penalty to stabilize the simultaneous training of image synthesis and classification. Through experiments on images of varying resolutions and complexity, we demonstrate that AC-GANs show promise in image classification, achieving competitive performance with standard CNNs. These methods can be employed as an 'all-in-one' framework with particular utility in the absence of large amounts of training data.