13
IRUS Total
Downloads
  Altmetric

3D high-resolution cardiac segmentation reconstruction from 2D views using conditional variational autoencoders

File Description SizeFormat 
1902.11000v1.pdfAccepted version2.31 MBAdobe PDFView/Open
Title: 3D high-resolution cardiac segmentation reconstruction from 2D views using conditional variational autoencoders
Authors: Biffi, C
Cerrolaza, JJ
Tarroni, G
De Marvao, A
Cook, SA
O'Regan, DP
Rueckert, D
Item Type: Conference Paper
Abstract: Accurate segmentation of heart structures imaged by cardiac MR is key for the quantitative analysis of pathology. High-resolution 3D MR sequences enable whole-heart structural imaging but are time-consuming, expensive to acquire and they often require long breath holds that are not suitable for patients. Consequently, multiplanar breath-hold 2D cines sequences are standard practice but are disadvantaged by lack of whole-heart coverage and low through-plane resolution. To address this, we propose a conditional variational autoencoder architecture able to learn a generative model of 3D high-resolution left ventricular (LV) segmentations which is conditioned on three 2D LV segmentations of one short-axis and two long-axis images. By only employing these three 2D segmentations, our model can efficiently reconstruct the 3D high-resolution LV segmentation of a subject. When evaluated on 400 unseen healthy volunteers, our model yielded an average Dice score of 87.92 ± 0.15 and outperformed competing architectures (TL-net, Dice score = 82.60 ± 0.23, p = 2.2 · 10 -16 ).
Issue Date: 11-Jul-2019
Date of Acceptance: 1-Apr-2019
URI: http://hdl.handle.net/10044/1/73720
DOI: https://doi.org/10.1109/ISBI.2019.8759328
ISSN: 1945-7928
Publisher: IEEE
Start Page: 1643
End Page: 1646
Journal / Book Title: 2019 IEEE 16TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI 2019)
Copyright Statement: ©2019 IEEE.
Sponsor/Funder: Imperial College Healthcare NHS Trust- BRC Funding
British Heart Foundation
Imperial College Healthcare NHS Trust- BRC Funding
Funder's Grant Number: RDC04
NH/17/1/32725
RDB02
Conference Name: 16th IEEE International Symposium on Biomedical Imaging (ISBI)
Keywords: Cardiac MR
Variational Autoencoder
3D Segmentation Reconstruction
Deep Learning
MASS
Cardiac MR
Variational Autoencoder
3D Segmentation Reconstruction
Deep Learning
MASS
cs.CV
cs.CV
Publication Status: Published
Start Date: 2019-04-08
Finish Date: 2019-04-11
Conference Place: Venice, ITALY
Online Publication Date: 2019-07-11
Appears in Collections:Computing
Institute of Clinical Sciences