Analysis of Heart Sound for Automated Diagnosis of Cardiac Disorders

Abstract

Body auscultation one of the main common clinical diagnostic technique for assessment of cardiac disorders. This clinical method is cheap and effective and requires professional doctors to understand and interpret the heart sound for diagnosis of cardiac diseases. This paper presents basic analysis and comparison of heart sounds recorded from healthy and unhealthy subjects for automated screening of cardiac disorders. Some characteristics of heart sounds such as probability distribution of amplitude and frequency contents are quantified using signal processing to automate the diagnosis of cardiac disorders. Experiments were carried on the Heart sounds database created under National Institute of Health (National Center for Research Resources) and results are encouraging. The experiment results indicates enough discrimination between heart sounds from healthy and unhealthy heart sounds for automated screening of cardiac disorders using signal processing.