Detection and classification of systolic murmur using a neural network

Abstract

Introduces a method for the automatic classification of the phonocardiogram. The method utilizes a multilayer perceptron for detecting the presence of the systolic murmur and a spectral analysis technique for separating innocent murmur from pathological murmur. The effectiveness of the method was confirmed by applying the method to data collected from a nationwide phonocardiogram screening study of Japanese elementary school children.