A continuous wavelet transform based method for detecting heart valve disorders using phonocardiograph signals

Abstract

The features extracted from the phonocardiograph (PCG) signals are commonly used for the classification of the heart valve disorders. This paper presents a new method based on features derived from the continuous wavelet transform (CWT) for classification of the PCG signals. In the proposed method, the PCG signal is segmented with enhanced modified p-spectrum based technique. The features are extracted from the segmented heart beat cycles using the CWT method. The two dimensional discrete cosine transform (2D-DCT) has been applied to the CWT coefficients in order to reduce the redundant information. The adaptively selected 2D-DCT coefficients have been used as input feature set for the least square support vector machine (LS-SVM) classifier. The experimental results are included in order to show the effectiveness of the proposed method for classification of heart valve disorders. © 2012 Springer-Verlag.