Predictions of blood flow varations based on artificial neural network and doppler signal

Abstract

In this approach a surrogate model based on Artificial Neural Networks (ANN), Doppler signal and Finite Element Method (FEM) is developed aiming to investigate the carotid arterial conditions. Doppler measurements of blood flow velocities are used as input values in Artificial Neural Network developments based on supervised learning. The network is trained using a Genetic Algorithm (GA) supported by an elitist strategy to obtain the optimal ANN topology. The ANN is cross validated and tested with carotid arterial Doppler signal and FEM analysis. The use of this ANN surrogate model for classification of carotid arterial diagnosis problems is proposed.