Heart murmur detection using cross match technique for primary care center

Abstract

Background: This study proposes cross-match technique to detect the presence of heart murmur in phonocardiogram. The sample data were recorded using an electronic stethoscope from real patient who suffers from structural heart disease such as Mitral Stenosis/Regurgitation and Aortic Stenosis/Regurgitation. The disease is confirmed by cardiologist with the support from an echocardiogram machine. Method: The data were segmented into cardiac cycles manually. Each cycle is carefully observed to ensure that only clean data is accepted for the experiment. The features were extracted using Instantaneous Energy and Frequency (IEFE) method. The selected features reflect the sound originated from the structural defects or physical malfunction of the heart mechanics during the blood pumping activities. Cross-match technique was then applied to train the input features to build the model barcode.