On the surface EMG signal reconstruction using blind source separation

In biomedical signal processing, many sources are often mixed as a form of measured signal. The goal is usually to extract and analyze one or several of them separately. In the multichannel measurements, several Blind Source Separation (BSS) techniques are available for decomposing the signal into its components. In this paper, a novel method is presented for the reconstruction of individual muscle source signals from simulated surface Elec-tromyography (s-EMG) array recordings. This method is based on BSS in a Bayesian model selection framework. Specifically, it is relies on an efficient wavelet spectral matching separating algorithm. Our concept is evaluated on theoretical decomposition and is confirmed by simulated signals.