Speaker: Hao Gao, Post-doctoral researcher first at Strathclyde University, later at the University of Glasgow as a research fellow in the School of Mathematics and Statistics. He is the key researcher and co-investigator of the SofTMech centre, one of five mathematics centre across UK. Currently, he is a lecturer in applied mathematics and (co)-leading several projects funded by EPSRC and BHF.

Date: July 6, 2023

Time: 15:00-16:00

Location: Tencent Meeting

Sponsor: School of Mathematics, Shandong University

Abstract:

The digital twin of a human heart, also known as a subject-specific digital model, built upon in vivo clinical measurements, has the potential to provide effective and consistent risk-stratification of patients. However, personalisation of these models can be challenging, particularly because different patients have different sets of biophysical parameters that can affect the quantities of interest, and it further requires running thousands of simulations of the mathematical model, which can be computationally expensive because these models are based on complex, nonlinear equations that do not have closed-form solutions. In this talk, I will first introduce the statistical emulation of computationally expensive cardiac models using Gaussian process with fixed computational domain and varied domains, then Bayesian inference of patient-specific biophysical parameters from real measurements. I will further talk about our recent work on automatic heart geometry reconstruction directly from medical images using state-of-art convolutional neural networks. Finally, I will discuss Graph Neural networks in emulating partial differential equations in complex domains both data-driven and physics-informed approaches.

For more information, please visit:

https://www.view.sdu.edu.cn/info/1020/181571.htm