Wavelet transform and simplicity based heart murmur segmentation

This paper is aimed at the identification of the boundaries of murmur present in heart sound. Heart murmurs provide crucial diagnosis information for several heart diseases such as natural or prosthetic valve dysfunction and heart failure. In order to find the valuable information about abnormal heart behavior, segmentation of the heart murmurs has to be performed. In this work we solve this problem using the wavelet decomposition- simplicity filter. In this algorithm, simplicity of the wavelet decomposed heart sound signal is measured and adaptively thresholded in order to discriminate S1/S2 sounds from murmurs. The method has been tested with stenosis and regurgitation (aortic, mitral, pulmonary, tricuspid) heart sounds and 89.10% sensitivity and 95.50 % specificity have been achieved.