Algorithm for Presumptive Diagnosis of Cardiac Murmurs over Auscultated Heart Signals

Abstract

The development of innovative cardiac assessment techniques has optimized heart conditions diagnosis, however it has led to a reduction of auscultation practice and hence its effectiveness. Furthermore, in public health systems it is not possible to meet the demand of patients who require the use of these new technologies, so auscultation is key in the process of evaluating cardiac subjects. Based on this situation, this document describes a proposal focused on the analysis and processing of the auscultated heart signal in order to generate a presumptive diagnosis of heart murmurs. The evaluation of the presence of a valvular lesion consisted in the identification of abnormalities in the relationship between the period of contraction and relaxation of the cardiac cycle. During the experimental phase of this work, acceptable results of sensitivity and specificity were obtained, reason why as future work it is planned its development as a tool of low cost to optimize the detection of valvular diseases.