A Shift-Invariant Wavelet in Mune Pattern Recognition

Authors

  • Jillian Salvador Electrical and Computer Engineering, McMaster University
  • Hubert deBruin Electrical and Computer Engineering, McMaster University

Abstract

A system was designed to estimate the number of motor units (MUNE) in a superficial muscle using incremental stimulation of a motor nerve followed by classification of the collected M-waves. In earlier work we used the Fourier power coefficients as pattern classifiers. The work presented examines the shift-invariant wavelet transform as an alternative M- wave classifier. The shift-invariant wavelet transform pattern classifier is compared to classification with the traditional wavelet transform vectors. Data to test the two approaches was obtained from the thenar muscles of six healthy subjects. The results show that the shift- invariant wavelet transform compensates for latency shifting and is superior to the traditional wavelet transform in classifying M-waves with smaller intra- class variances.

Key words – motor unit number estimation, MUNE, electromyography, wavelet analysis, shift-invariant wavelet, motor unit action potentials 

Downloads

Published

2007-12-31

How to Cite

[1]
J. Salvador and H. deBruin, “A Shift-Invariant Wavelet in Mune Pattern Recognition”, CMBES Proc., vol. 30, no. 1, Dec. 2007.

Issue

Section

Academic