Convolutional Neural Network for Improvement of Heart Valve Disease Detection
Abstract
Valvular heart disease (VHD) encompasses a number of common cardiovascular conditions that account for a significant percentage of heart diseases. At present, the acoustic phenomena generated by the abnormal functioning of the heart valves can be recorded and digitized using electronic stethoscopes known as phonocardiographs. The analysis of the
phonocardiographic signals has made it possible to indicate that the normal and pathological records differ from each other in terms of both temporal and spectral characteristics. The present work describes the construction and implementation of a Deep Learning (DL) algorithm for the binary classification of normal and abnormal heart sounds. The performance of this approach reached an accuracy higher than 98 % and specificities in the ”Normal” class of up to 99 %.
phonocardiographic signals has made it possible to indicate that the normal and pathological records differ from each other in terms of both temporal and spectral characteristics. The present work describes the construction and implementation of a Deep Learning (DL) algorithm for the binary classification of normal and abnormal heart sounds. The performance of this approach reached an accuracy higher than 98 % and specificities in the ”Normal” class of up to 99 %.
Keywords
Artificial intelligence, deep neural network, phonocardiography, heart valve disease