IRUS Total

Realistic adversarial data augmentation for MR image segmentation

File Description SizeFormat 
2006.13322v1.pdfWorking paper3.67 MBAdobe PDFView/Open
Title: Realistic adversarial data augmentation for MR image segmentation
Authors: Chen, C
Qin, C
Qiu, H
Ouyang, C
Wang, S
Chen, L
Tarroni, G
Bai, W
Rueckert, D
Item Type: Working Paper
Abstract: Neural network-based approaches can achieve high accuracy in various medical image segmentation tasks. However, they generally require large labelled datasets for supervised learning. Acquiring and manually labelling a large medical dataset is expensive and sometimes impractical due to data sharing and privacy issues. In this work, we propose an adversarial data augmentation method for training neural networks for medical image segmentation. Instead of generating pixel-wise adversarial attacks, our model generates plausible and realistic signal corruptions, which models the intensity inhomogeneities caused by a common type of artefacts in MR imaging: bias field. The proposed method does not rely on generative networks, and can be used as a plug-in module for general segmentation networks in both supervised and semi-supervised learning. Using cardiac MR imaging we show that such an approach can improve the generalization ability and robustness of models as well as provide significant improvements in low-data scenarios.
Issue Date: 23-Jun-2020
URI: http://hdl.handle.net/10044/1/80458
Publisher: arXiv
Copyright Statement: © 2020 The Author(s)
Sponsor/Funder: Engineering & Physical Science Research Council (EPSRC)
Funder's Grant Number: EP/P001009/1
Keywords: eess.IV
Notes: 13 pages. This paper is accepted to MICCAI 2020
Publication Status: Published
Appears in Collections:Computing
Faculty of Medicine
Department of Brain Sciences
Faculty of Engineering