A Comparative Analysis of Denoising PCG Signals using Adaptive Wiener-Kalman Filtering

Cardiovascular disease (CVD) is the group of disorder occurs on heart and blood vessel, and they are coronary heart disease, cerebrovascular disease, rheumatic heart disease, etc. Particularly people under the age of 70 are mostly affected by this disease. Earlier cardiologist analyse phonocardiography (PCG) signals to know the condition of heart. Because of the undesirable noise present in PCG signal makes complication in diagnosis by experts. This study analyse and denoising heart sound using five different techniques namely Short-Time Fourier Transform (STFT), Empirical Mode Decomposition (EMD), Recurrent Neural Network (RNN), weiner filter (WF), adaptive Wiener-Kalman filtering (AWKF). There are Lot of noise removal techniques available as discussed in the literature, failed to achieve higher performance in denoising audio signal. The developed denoising techniques were applied on various pathological heart sounds and on heart sound recorded in a noisy environment. The AWKF filtering has great impact on removing distortion and interference present on the signal. This method reserves the major characteristics of heart beat sound. The experimental results showed that AWKF acquired higher SNR, PSNR, lower RMSE and great denoising effect.