Phonocardiogram Classification using Deep Learning

Abstract

According to World Health Organization 17.7 million people die because of the cardiovascular diseases. So, to be able to mitigate the risk of heart related problems early detection of condition of heart may save millions of lives worldwide. We have built a solution which would classify the heart sound into normal and abnormal. Along with that there is a mobile application so that it is easy for the Cardiologists, Doctors, and other non-technical people to check the Phonocardiogram(PCG). Our system is intended to mitigate the risk of error of manual auscultation which is considered to be error prone. The system works as a computer-aided tool which is cost effective and time efficient. Thereby, aid the Physicians and Cardiologists. The system provides an overview of the person's heart and its condition.