Noninvasive detection of coronary artery disease

Recently, the authors focused on the application of the neural networks to improve the diagnostic capability of the acoustical approach. In order to improve the diagnostic ability for mis-diagnosed patients, the combination of the first four moments (mean, variance, skewness, kurtosis) of the extrema of the coefficients of wavelet transform applied to the diastolic heart sounds associated with coronary artery disease, as well as physical examination parameters, were used as the input pattern to the neural networks. The wavelet transform was chosen, since it is free from assumptions concerning the characteristics of the signal. Finally, using their nonlinear and multilayered architecture, fuzzy neural networks were applied to the diastolic heart sounds produced by coronary stenoses in order to capture fully all relevant information related to the patients' disease states.>