EMG signal denoising via Bayesian wavelet shrinkage based on GARCH modeling

In this paper, we introduce a novel noise suppression method for electromyography (EMG) signals, based on statistical modeling of wavelet coefficients. First, we demonstrate that Generalized Autoregressive Conditional Heteroscedasticity (GARCH) effect exists in wavelet coefficients of EMG signals. Then, we use GARCH model for these coefficients. In consequence, we introduce a maximum a-posteriori (MAP) estimator, based on GARCH modeling, for estimating the clean wavelet coefficients. To evaluate the performance of GARCH based method in noise suppression, we compare our proposed method with other wavelet based denoising methods and we verify the performance improvement in utilizing the new strategy.