Evaluation of Pattern Recognition in Myoelectric Signal Using Netlab GLM

Abstract

The myoelectric signal that is collected from the surface of the skin can be used to construct rehabilitation systems for people who have suffered some trauma or who were born with some form of malformation. This signal is used to feed classifiers that can tell with some degree of distinction which movement each signal belongs to. Among the approaches used for this task are the use of artificial neural networks (ANN), multi-layer perceptron (MLP), linear discriminant models (LDA), among others. In this study a approach to pattern recognition called Netlab GLM that has two optimized methods for network training is evaluated in different situations. The classical algorithm LDA is used as a criterion of comparison.