Overlapping group sparse denosing: a good choice for noise removal from EMG signal in intermittent masseter muscle activity

Abstract

Biological signals are often impregnated with a variety of noises. Then noise removal for precise processing is very important. The aim of this study was to test a method called Overlapping Group Sparse Denoising performance to remove noises from electroencephalography signals of sequential masseter activity. Overlapping group sparse denoising method was studied on EMG signals obtained from the masseter muscle. The Electromyography (EMG) signals obtained from three groups of control healthy participants, migraine without aura and tension headache patients. Four metrics (MSE, MAE, SNR and PSNR) calculated for analysing the method performance in the denoising EMG signals. The results indicated that using mentioned method was successful. The method can be helpful for denoising signals with intervals of clustering activations and deactivation like sound signals.