Feeding the zombies: Synthesizing brain volumes using a 3D progressive growing GAN

TL;DR

This paper demonstrates that a 3D progressive growing GAN can generate realistic synthetic brain MRI volumes, addressing data scarcity in neuroimaging for training deep learning models.

Contribution

It introduces the application of a 3D progressive growing GAN to synthesize neuroimaging data, a novel approach in medical imaging research.

Findings

Successfully generated realistic 3D brain MR volumes

Addresses data scarcity in neuroimaging datasets

Paves the way for improved deep learning training with synthetic data

Abstract

Deep learning requires large datasets for training (convolutional) networks with millions of parameters. In neuroimaging, there are few open datasets with more than 100 subjects, which makes it difficult to, for example, train a classifier to discriminate controls from diseased persons. Generative adversarial networks (GANs) can be used to synthesize data, but virtually all research is focused on 2D images. In medical imaging, and especially in neuroimaging, most datasets are 3D or 4D. Here we therefore present preliminary results showing that a 3D progressive growing GAN can be used to synthesize MR brain volumes.