TY - GEN
T1 - Semi-supervised learning for network-based cardiac MR image segmentation
AU - Bai, Wenjia
AU - Oktay, Ozan
AU - Sinclair, Matthew
AU - Suzuki, Hideaki
AU - Rajchl, Martin
AU - Tarroni, Giacomo
AU - Glocker, Ben
AU - King, Andrew
AU - Matthews, Paul M.
AU - Rueckert, Daniel
N1 - Funding Information:
Acknowledgements. This research has been conducted using the UK Biobank Resource under Application Number 18545. This work is supported by EPSRC programme Grant (EP/P001009/1). H.S. is supported by a Research Fellowship from the Uehara Memorial Foundation. P.M.M. gratefully acknowledges support from the Imperial College Healthcare Trust Biomedical Research Centre, the EPSRC Centre for Mathematics in Precision Healthcare and the MRC.
Publisher Copyright:
© Springer International Publishing AG 2017.
PY - 2017
Y1 - 2017
N2 - Training a fully convolutional network for pixel-wise (or voxel-wise) image segmentation normally requires a large number of training images with corresponding ground truth label maps. However, it is a challenge to obtain such a large training set in the medical imaging domain, where expert annotations are time-consuming and difficult to obtain. In this paper, we propose a semi-supervised learning approach, in which a segmentation network is trained from both labelled and unlabelled data. The network parameters and the segmentations for the unlabelled data are alternately updated. We evaluate the method for short-axis cardiac MR image segmentation and it has demonstrated a high performance, outperforming a baseline supervised method. The mean Dice overlap metric is 0.92 for the left ventricular cavity, 0.85 for the myocardium and 0.89 for the right ventricular cavity. It also outperforms a state-of-the-art multi-atlas segmentation method by a large margin and the speed is substantially faster.
AB - Training a fully convolutional network for pixel-wise (or voxel-wise) image segmentation normally requires a large number of training images with corresponding ground truth label maps. However, it is a challenge to obtain such a large training set in the medical imaging domain, where expert annotations are time-consuming and difficult to obtain. In this paper, we propose a semi-supervised learning approach, in which a segmentation network is trained from both labelled and unlabelled data. The network parameters and the segmentations for the unlabelled data are alternately updated. We evaluate the method for short-axis cardiac MR image segmentation and it has demonstrated a high performance, outperforming a baseline supervised method. The mean Dice overlap metric is 0.92 for the left ventricular cavity, 0.85 for the myocardium and 0.89 for the right ventricular cavity. It also outperforms a state-of-the-art multi-atlas segmentation method by a large margin and the speed is substantially faster.
UR - http://www.scopus.com/inward/record.url?scp=85029504302&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85029504302&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-66185-8_29
DO - 10.1007/978-3-319-66185-8_29
M3 - Conference contribution
AN - SCOPUS:85029504302
SN - 9783319661841
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 253
EP - 260
BT - Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings
A2 - Jannin, Pierre
A2 - Duchesne, Simon
A2 - Descoteaux, Maxime
A2 - Franz, Alfred
A2 - Collins, D. Louis
A2 - Maier-Hein, Lena
PB - Springer Verlag
T2 - 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017
Y2 - 11 September 2017 through 13 September 2017
ER -