Towards disease-aware image editing of chest X-rays

TL;DR

This paper demonstrates the feasibility of using GANs, specifically StyleGAN, for disease-aware editing of chest X-ray images, enabling manipulation of medical conditions like cardiomegaly.

Contribution

It introduces a method to invert GAN-generated images and manipulate disease features in chest X-rays, advancing AI applications in medical image editing.

Findings

StyleGAN can generate realistic chest X-rays.

Latent space manipulation can alter medical conditions.

Organ-level reconstructions of real X-rays are possible.

Abstract

Disease-aware image editing by means of generative adversarial networks (GANs) constitutes a promising avenue for advancing the use of AI in the healthcare sector. Here, we present a proof of concept of this idea. While GAN-based techniques have been successful in generating and manipulating natural images, their application to the medical domain, however, is still in its infancy. Working with the CheXpert data set, we show that StyleGAN can be trained to generate realistic chest X-rays. Inspired by the Cyclic Reverse Generator (CRG) framework, we train an encoder that allows for faithfully inverting the generator on synthetic X-rays and provides organ-level reconstructions of real ones. Employing a guided manipulation of latent codes, we confer the medical condition of cardiomegaly (increased heart size) onto real X-rays from healthy patients. This work was presented in the Medical Imaging meets Neurips Workshop 2020, which was held as part of the 34th Conference on Neural Information Processing Systems (NeurIPS 2020) in Vancouver, Canada