Research on High Resolution Analysis of Heart Sound and Identification Method for Valvular Heart Disease Examination

点击翻译按钮获取中文摘要

In this study, we first developed a prototype of auscultation cloth which is easy to wear and to collect good quality heart sound even by general users. The heart sounds both from the normal persons and heart disease patients have been collected by using the auscultation wear. The collected sound signals including their feature parameters obtained by analysis have been made in the heart disease sound database. To improve the resolution both in time and frequency domains analysis, a novel time-frequency analysis, called as Frequency Slice Wavelet Transform(FSWT), was proposed. By using this method, it is possible to extract more accurate heart sound feature parameters both in the frequency and time domains. By applying the data clustering method(SVM) to these parameters, the results shown that the heart disease can be identified with high accuracy.