Machine learning augmented reduced-order models for FFR-prediction

DOI10.1016/j.cma.2021.113892zbMath1506.65025OpenAlexW3169457531MaRDI QIDQ2237421

Fredrik E. Fossan, Jacob T. Sturdy, Leif R. Hellevik, Rune Wiseth, Anders T. Bråten, Arve Jørgensen, Lucas O. Müller

Publication date: 27 October 2021

Published in: Computer Methods in Applied Mechanics and Engineering (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1016/j.cma.2021.113892

zbMATH Keywords

reduced-order modelingphysics-informed neural networkscomputational FFR

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Biomedical imaging and signal processing (92C55) Probabilistic methods, stochastic differential equations (65C99)

Related Items

Finite element coupled positive definite deep neural networks mechanics system for constitutive modeling of composites, Patient-specific modeling of blood flow in the coronary arteries, Prediction of fractional flow reserve based on reduced-order cardiovascular model

Uses Software

• FEniCS
• TensorFlow
• Keras
• Adam

Cites Work

• Automated solution of differential equations by the finite element method. The FEniCS book
• Machine learning in cardiovascular flows modeling: predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
• Machine learning of multiscale active force generation models for the efficient simulation of cardiac electromechanics
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Impact of geometric uncertainty on hemodynamic simulations using machine learning