Frequency Extraction of Phonocardiogram Signal using Fourier Transform

This article presents Fourier Transform application to extract features of Phonocardiogram Signals into its frequency components. Data was taken from Physionet Dataset of Phonocardiogram which comprises of normal and abnormal heart condition. Raw data was preprocessed using time clipping of 2 seconds at certain area that contains less noise. A lowpass filter was applied to denoise the raw signals. Experiments show the PCG of normal hearts has a dominant frequency of 50Hz to 150Hz, with the subdominant frequencies of 450 Hz to 650 Hz. The subdominant frequency of the normal hearts sometimes show anomaly with more amplitude compared to the dominant frequency.