EMG artifacts removal during electrical stimulation, a CWT based technique

Maxime Yochum (1) , Toufik Bakir (1) , Stéphane Binczak (1) , Romuald Lepers (2)

1 Le2i - Laboratoire Electronique, Informatique et Image [UMR6306]
2 CAPS - Cognition, Action, et Plasticité Sensorimotrice [Dijon - U1093]

A technique of artifacts removal based on the continuous wavelet transform is presented. It uses common mother wavelets to find the temporal localization of stimulation artifacts on electromyogram (EMG) signal during an electrically evoked contraction of a muscle. This method can be used with standard stimulation pulse waveforms like monophasics or biphasics ones. It uses a histogram representation to find the best threshold to apply on the CWT domain. The algotithm is presented with Haar wavelet and then it is used with common wavelet famillies such as Daubechies or Symlets.

Mots clés: Electromyogram; stimulation artifacts; source separation; wavelet; histogram representation

Domaines: Traitement du signal et de l'image [eess.SP]

Maxime Yochum, Toufik Bakir, Stéphane Binczak, Romuald Lepers. EMG artifacts removal during electrical stimulation, a CWT based technique. IEEE Region 10 Symposium, Apr 2014, Kuala Lumpur, Malaysia. ⟨10.1109/TENCONSpring.2014.6863012⟩. ⟨hal-01328483⟩