Automatic discrimination between heart sounds and murmurs using parametric models

• Published in: Proceedings of the 1988 IEEE International Conference on Systems, Man, and Cybernetics Vol. 1; pp. 147 - 150
• Main Authors: Mohamed, A.S.A., Raafat, H.M.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 1988
• Subjects: Blood; Cardiac disease; Cardiology; Computer science; Ear; Frequency; Heart valves; Parametric statistics; Solids; Stethoscope
• ISBN: 7800030393; 9787800030390
• DOI: 10.1109/ICSMC.1988.754260

Most of the cardiovascular sounds are produced within or about the heart and great vessels and are transmitted through liquid and solid media to the chest wall. During auscultation of the heart, we receive air transmitted sounds through the stethoscope which is limited by the sensitivity of the ear. Present methods for interpretation of heart sounds and murmurs[HSAM]are qualitative and crude depending mainly on experience. Therefore, there is a need for more sensitive and quantitative methods which would permit precise measures for diagnosis. This paper develops a mathematical model to describe the HSAM by a finite number of parameters. The auto regressive model[AR] is selected to represent the HSAM at the principal areas of cardiac auscultation and for different heart diseases. Validation of the model results is based on the minimum mean squared deviation between the signal spectrum obtained by FFT and the estimated spectrum obtained from the AR model. A data base from the simulated HSAM is created and discrimination among the measured HSAM from subjects with different heart diseases is performed. Based on Bayes rule, the discrimination between heart sounds and murmurs is performed.