22 October at 16:30 MSK
Modern blood flow models can be used to estimate a variety of important diagnostic indices. One of the problems that hinders their applicability is parameter identification. Some parameters are expensive to obtain, but they are important for model personalization. We propose a method to estimate some of the hard-to-get parameters. This method is based on a synthetic data base that was generated with the help of a separate blood flow model of systemic circulation and consists of 4374 virtual patients. We use neural network trained on synthetic data to estimate some of the important indices from patient’s age, blood pressure, height, weight, heart rate and other easy-to-obtain parameters.
Gamilov Timur, senior researcher, laboratory of mathematical modelling, Sechenov University.
Topic: Blood flow model parameters estimation with the help of synthetic database.