Fast Fully Automatic Segmentation of the Severely Abnormal Human Right Ventricle from Cardiovascular Magnetic Resonance Images Using a Multi-Scale 3D Convolutional Neural Network

Giannakidis A, Kamnitsas K, Spadotto V, Keegan J, Smith G, Glocker B, Rueckert D, Ernst S, Gatzoulis MA, Pennell DJ, Babu-Narayan S, Firmin DN (2017)

Publication Type: Conference contribution

Publication year: 2017

Publisher: Institute of Electrical and Electronics Engineers Inc.

Pages Range: 42-46

Conference Proceedings Title: Proceedings - 12th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2016

Event location: Naples, ITA

ISBN: 9781509056989

DOI: 10.1109/SITIS.2016.16

Abstract

Cardiac magnetic resonance (CMR) is regarded as the reference examination for cardiac morphology in tetralogy of Fallot (ToF) patients allowing images of high spatial resolution and high contrast. The detailed knowledge of the right ventricular anatomy is critical in ToF management. The segmentation of the right ventricle (RV) in CMR images from ToF patients is a challenging task due to the high shape and image quality variability. In this paper we propose a fully automatic deep learning-based framework to segment the RV from CMR anatomical images of the whole heart. We adopt a 3D multi-scale deep convolutional neural network to identify pixels that belong to the RV. Our robust segmentation framework was tested on 26 ToF patients achieving a Dice similarity coefficient of 0.8281±0.1010 with reference to manual annotations performed by expert cardiologists. The proposed technique is also computationally efficient, which may further facilitate its adoption in the clinical routine.

Involved external institutions

Imperial College London / The Imperial College of Science, Technology and Medicine GB United Kingdom (GB) Royal Brompton Hospital GB United Kingdom (GB)

How to cite

APA:

Giannakidis, A., Kamnitsas, K., Spadotto, V., Keegan, J., Smith, G., Glocker, B.,... Firmin, D.N. (2017). Fast Fully Automatic Segmentation of the Severely Abnormal Human Right Ventricle from Cardiovascular Magnetic Resonance Images Using a Multi-Scale 3D Convolutional Neural Network. In Giuseppe De Pietro, Albert Dipanda, Richard Chbeir, Luigi Gallo, Kokou Yetongnon (Eds.), Proceedings - 12th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2016 (pp. 42-46). Naples, ITA: Institute of Electrical and Electronics Engineers Inc..

MLA:

Giannakidis, Archontis, et al. "Fast Fully Automatic Segmentation of the Severely Abnormal Human Right Ventricle from Cardiovascular Magnetic Resonance Images Using a Multi-Scale 3D Convolutional Neural Network." Proceedings of the 12th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2016, Naples, ITA Ed. Giuseppe De Pietro, Albert Dipanda, Richard Chbeir, Luigi Gallo, Kokou Yetongnon, Institute of Electrical and Electronics Engineers Inc., 2017. 42-46.

BibTeX: Download