Heart Sound is one of the oldest means for assessing the function of its valves. It helps, together with Echocardiograms and Electrocardiographs, giving a clear and proper diagnostic of several diseases. In this paper artificial neural networks are used to classify several valves-related heart disorders. A library of heart sound files, recorded via the traditional Stethoscope, are used to extract relevant features using several signal processing tools e.g Discrete Wavelet Transfer (DWT), Fast Fourier Transform (FFT) and Linear Prediction Coding (LPC). The achieved recognition rates were around 95.7%.
展开▼