MaxStyle: Adversarial Style Composition for Robust Medical Image Segmentation

Chen Chen, Zeju Li, Cheng Ouyang, Matthew Sinclair, Wenjia Bai, Daniel Rueckert

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

10 Scopus citations

Abstract

Convolutional neural networks (CNNs) have achieved remarkable segmentation accuracy on benchmark datasets where training and test sets are from the same domain, yet their performance can degrade significantly on unseen domains, which hinders the deployment of CNNs in many clinical scenarios. Most existing works improve model out-of-domain (OOD) robustness by collecting multi-domain datasets for training, which is expensive and may not always be feasible due to privacy and logistical issues. In this work, we focus on improving model robustness using a single-domain dataset only. We propose a novel data augmentation framework called MaxStyle, which maximizes the effectiveness of style augmentation for model OOD performance. It attaches an auxiliary style-augmented image decoder to a segmentation network for robust feature learning and data augmentation. Importantly, MaxStyle augments data with improved image style diversity and hardness, by expanding the style space with noise and searching for the worst-case style composition of latent features via adversarial training. With extensive experiments on multiple public cardiac and prostate MR datasets, we demonstrate that MaxStyle leads to significantly improved out-of-distribution robustness against unseen corruptions as well as common distribution shifts across multiple, different, unseen sites and unknown image sequences under both low- and high-training data settings. The code can be found at https://github.com/cherise215/MaxStyle.

Original languageEnglish
Title of host publicationMedical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings
EditorsLinwei Wang, Qi Dou, P. Thomas Fletcher, Stefanie Speidel, Shuo Li
PublisherSpringer Science and Business Media Deutschland GmbH
Pages151-161
Number of pages11
ISBN (Print)9783031164422
DOIs
StatePublished - 2022
Externally publishedYes
Event25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Singapore, Singapore
Duration: 18 Sep 202222 Sep 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13435 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
Country/TerritorySingapore
CitySingapore
Period18/09/2222/09/22

Fingerprint

Dive into the research topics of 'MaxStyle: Adversarial Style Composition for Robust Medical Image Segmentation'. Together they form a unique fingerprint.

Cite this