Imperial College London
Portrait Picture

Tom Wong

Faculty of Medicine, National Heart & Lung Institute

Reader in Cardiology
 
 
 
//

Contact

 

+44 (0)20 7351 8619tom.wong

 
 
//

Assistant

 

Dr Vias Markides +44 (0)20 7351 8619

 
//

Location

 

Chelsea WingRoyal Brompton Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Chen:2022:10.1109/JBHI.2021.3077469,
author = {Chen, J and Yang, G and Khan, H and Zhang, H and Zhang, Y and Zhao, S and Mohiaddin, R and Wong, T and Firmin, D and Keegan, J},
doi = {10.1109/JBHI.2021.3077469},
journal = {IEEE Journal of Biomedical and Health Informatics},
pages = {103--114},
title = {JAS-GAN: generative adversarial network based joint atrium and scar segmentations on unbalanced atrial targets},
url = {http://dx.doi.org/10.1109/JBHI.2021.3077469},
volume = {26},
year = {2022}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - Automated and accurate segmentation of the left atrium (LA) and atrial scars from late gadolinium-enhanced cardiac magnetic resonance (LGE CMR) images are in high demand for quantifying atrial scars. The previous quantification of atrial scars relies on a two-phase segmentation for LA and atrial scars due to their large volume difference (unbalanced atrial targets). In this paper, we propose an inter-cascade generative adversarial network, namely JAS-GAN, to segment the unbalanced atrial targets from LGE CMR images automatically and accurately in an end-to-end way. Firstly, JAS-GAN investigates an adaptive attention cascade to automatically correlate the segmentation tasks of the unbalanced atrial targets. The adaptive attention cascade mainly models the inclusion relationship of the two unbalanced atrial targets, where the estimated LA acts as the attention map to adaptively focus on the small atrial scars roughly. Then, an adversarial regularization is applied to the segmentation tasks of the unbalanced atrial targets for making a consistent optimization. It mainly forces the estimated joint distribution of LA and atrial scars to match the real ones. We evaluated the performance of our JAS-GAN on a 3D LGE CMR dataset with 192 scans. Compared with state-of-the-art methods, our proposed approach yielded better segmentation performance (Average Dice Similarity Coefficient (DSC) values of 0.946 and 0.821 for LA and atrial scars, respectively), which indicated the effectiveness of our proposed approach for segmenting unbalanced atrial targets.
AU  - Chen,J
AU  - Yang,G
AU  - Khan,H
AU  - Zhang,H
AU  - Zhang,Y
AU  - Zhao,S
AU  - Mohiaddin,R
AU  - Wong,T
AU  - Firmin,D
AU  - Keegan,J
DO  - 10.1109/JBHI.2021.3077469
EP  - 114
PY  - 2022///
SN  - 2168-2194
SP  - 103
TI  - JAS-GAN: generative adversarial network based joint atrium and scar segmentations on unbalanced atrial targets
T2  - IEEE Journal of Biomedical and Health Informatics
UR  - http://dx.doi.org/10.1109/JBHI.2021.3077469
UR  - https://journals.lww.com/pain/Abstract/9000/Pragmatic_trials_of_pain_therapies__a_systematic.98036.aspx
UR  - http://hdl.handle.net/10044/1/89713
VL  - 26
ER  -
Download