Imperial College London
Portrait Picture

DrWenjiaBai

Faculty of Medicine, Department of Brain Sciences

Senior Lecturer
 
 
 
//

Contact

 

+44 (0)20 7594 8291w.bai Website

 
 
//

Location

 

Room 212, Data Science InstituteWilliam Penney LaboratorySouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@inproceedings{Qin:2020:10.1007/978-3-030-59716-0_29,
author = {Qin, C and Wang, S and Chen, C and Qiu, H and Bai, W and Rueckert, D},
doi = {10.1007/978-3-030-59716-0_29},
pages = {296--306},
publisher = {Springer International Publishing},
title = {Biomechanics-informed neural networks for myocardial motion tracking in MRI},
url = {http://dx.doi.org/10.1007/978-3-030-59716-0_29},
year = {2020}
}
Download

RIS format (EndNote, RefMan)

TY  - CPAPER
AB  - Image registration is an ill-posed inverse problem which often requires regularisation on the solution space. In contrast to most of the current approaches which impose explicit regularisation terms such as smoothness, in this paper we propose a novel method that can implicitly learn biomechanics-informed regularisation. Such an approach can incorporate application-specific prior knowledge into deep learning based registration. Particularly, the proposed biomechanics-informed regularisation leverages a variational autoencoder (VAE) to learn a manifold for biomechanically plausible deformations and to implicitly capture their underlying properties via reconstructing biomechanical simulations. The learnt VAE regulariser then can be coupled with any deep learning based registration network to regularise the solution space to be biomechanically plausible. The proposed method is validated in the context of myocardial motion tracking on 2D stacks of cardiac MRI data from two different datasets. The results show that it can achieve better performance against other competing methods in terms of motion tracking accuracy and has the ability to learn biomechanical properties such as incompressibility and strains. The method has also been shown to have better generalisability to unseen domains compared with commonly used L2 regularisation schemes.
AU  - Qin,C
AU  - Wang,S
AU  - Chen,C
AU  - Qiu,H
AU  - Bai,W
AU  - Rueckert,D
DO  - 10.1007/978-3-030-59716-0_29
EP  - 306
PB  - Springer International Publishing
PY  - 2020///
SN  - 0302-9743
SP  - 296
TI  - Biomechanics-informed neural networks for myocardial motion tracking in MRI
UR  - http://dx.doi.org/10.1007/978-3-030-59716-0_29
UR  - https://link.springer.com/chapter/10.1007%2F978-3-030-59716-0_29
ER  -
Download