Alzheimer Brain Imaging Dataset Augmentation Using Wasserstein Generative Adversarial Network

Abstract

Deep learning models have evolved to be very efficient and robust for several computer vision applications. To harness the benefits of state-of-the-art deep networks in the realm of disease detection and prediction, it is imperative that high-quality datasets be made available for the models to train on. This work recognizes the dearth of training data (both in terms of quality and quantity of images) for using such networks for the detection of Alzheimer’s disease. It is proposed to employ a Wasserstein Generative Adversarial Network (WGAN) for generating synthetic images for augmentation of an existing Alzheimer brain image dataset. It is shown that the proposed approach is indeed successful in generating high-quality images for inclusion in the Alzheimer image dataset potentially making the dataset more suited for training high-end models.