GANs for Medical Image Analysis

TL;DR

This review paper discusses how GANs are used in medical image analysis to improve tasks like de-noising, segmentation, and data synthesis, addressing data scarcity and enhancing image realism.

Contribution

It provides a comprehensive overview of recent GAN applications in medicine, analyzing their strengths, limitations, and future research directions.

Findings

GANs improve image quality and data augmentation in medical imaging

They enable realistic image synthesis for training and diagnosis

Current challenges include mode collapse and limited clinical validation

Abstract

Generative Adversarial Networks (GANs) and their extensions have carved open many exciting ways to tackle well known and challenging medical image analysis problems such as medical image de-noising, reconstruction, segmentation, data simulation, detection or classification. Furthermore, their ability to synthesize images at unprecedented levels of realism also gives hope that the chronic scarcity of labeled data in the medical field can be resolved with the help of these generative models. In this review paper, a broad overview of recent literature on GANs for medical applications is given, the shortcomings and opportunities of the proposed methods are thoroughly discussed and potential future work is elaborated. We review the most relevant papers published until the submission date. For quick access, important details such as the underlying method, datasets and performance are tabulated. An interactive visualization which categorizes all papers to keep the review alive, is available at http://livingreview.in.tum.de/GANs_for_Medical_Applications.