Cascaded triplanar autoencoder M-Net for fully automatic segmentation of left ventricle myocardial scar from three-dimensional late gadolinium-enhanced MR images

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

16 Scopus Citations
View graph of relations

Author(s)

  • Mingbo Zhao
  • Eranga Ukwatta
  • James A. White

Related Research Unit(s)

Detail(s)

Original languageEnglish
Pages (from-to)2582-2593
Journal / PublicationIEEE Journal of Biomedical and Health Informatics
Volume26
Issue number6
Online published25 Jan 2022
Publication statusPublished - Jun 2022

Abstract

While 3D three-dimensional (3D) late gadolinium-enhanced (LGE) magnetic resonance (MR) imaging provides good conspicuity of small myocardial lesions with short acquisition time, it poses a challenge for image analysis as a large number of axial images are required to be segmented. We developed a fully automatic convolutional neural network (CNN) called cascaded triplanar autoencoder M-Net (CTAEM-Net) to segment myocardial scar from 3D LGE MRI. Two sub-networks were cascaded to segment the left ventricle (LV) myocardium and then the scar within the presegmented LV myocardium. Each sub-network contains three autoencoder M-Nets (AEM-Nets) segmenting the axial, sagittal and coronal slices of the 3D LGE MR image, with the final segmentation produced by each sub-network determined by voting. The AEM-Net integrates three features in its design: (1) multi-scale inputs, (2) deep supervision and (3) multi-tasking. The multi-scale inputs allow consideration of the global and local features in segmentation. Deep supervision provides direct supervision to deeper layers and facilitates CNN convergence. Multi-task learning reduces overfitting in segmentation by acquiring additional information from autoencoder reconstruction, a task closely related to segmentation. The framework provides an accuracy of 86.43% and 90.18% for LV myocardium and scar segmentation, respectively, which are the highest among existing methods to our knowledge. The time required for CTAEM-Net to segment LV myocardium and the scar was 49.72 9.69 s and 120.25 23.18 s per MR volume. The accuracy and efficiency afforded by CTAEM-Net in scar segmentation will make possible future studies involving a large population. The generalizability of the framework was also demonstrated by its competitive performance in two publicly available datasets of different imaging modalities.

Research Area(s)

  • Convolutional neural networks, Image segmentation, Lesions, Magnetic resonance imaging, Myocardium, Task analysis, Three-dimensional displays, Deep learning, 3D late gadoliniumenhanced magnetic resonance (LGE MR) images, left ventricle myocardial scar segmentation

Citation Format(s)

Cascaded triplanar autoencoder M-Net for fully automatic segmentation of left ventricle myocardial scar from three-dimensional late gadolinium-enhanced MR images. / Lin, Mingquan; Jiang, Mingjie; Zhao, Mingbo et al.
In: IEEE Journal of Biomedical and Health Informatics, Vol. 26, No. 6, 06.2022, p. 2582-2593.

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review