Portable heart valve disease screening device using electronic stethoscope

Abstract

Heart sound analysis has been a popular topic of studies since a few decades ago. Most of the studies are done in PC platform since embedding the complex algorithm into a simple small device such as microcontroller board seems to be very difficult due to limited processing speed and memory. This study classifies normal and abnormal heart sound signal from four categories of Heart Valve Disease. An automated system that consists of segmentation, feature extraction and classification of the heart sound signal is developed in PC and hardware platforms. A multimedia board completed with a single board computer, audio codec and graphic LCD is used to make a portable heart valve disease screening device with electronic stethoscope as the input for the system. Both system recorded 96.3% specificity. However, the portable device has only 77.78% sensitivity and 87.04% accuracy compared to PC platform that have sensitivity and accuracy of more than 90%.