STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Image Processing & Pattern Recognition Progress

Pattern Recognition of Finger Movements of Two Channel based Surface EMG Signals

  1. A Goen
  2. D C Tiwari

Abstract

Surface EMG signal is widely used as a control signal for prosthetic device for the amputees or paralyzed able-bodied person with added advantage of being noninvasive and not painful. An ample amount of research on the prosthetic control of movements of forearm and hand is available in the literatures but few researches has been carried out for control of more dexterous individual and combined finger movements. The ability to precisely control a large number of individual and combined finger movements in a computationally efficient manner is very much required. This paper finds out the discrimination between individual and combined finger movements using surface myoelectric signals. We have SEMG datasets with two electrodes located on the human forearm for the classification of ten classes. Various feature sets were extracted, selected and projected in such a way to ensure that maximum separation exists between the different classes of finger movements. We have introduced two classifiers Modified kNN (MkNN) and SVM ensemble which are new in the field of prosthetic control. Also we have used two new projection techniques which have been rarely used in any of the literatures of prosthetic control. The set of practical results proved the feasibility of the proposed approach with mean classification accuracy greater than 98% in finger movement classification.Cite this ArticleGoen A, Tiwari DC. Pattern recognition of finger movements of two channel based surface EMG signals. Journal of Image Processing & Pattern Recognition Progress. 2016; 3(1): 7–14p.

Keywords

Support