EMG Classification Based On Features Reduction Using Fuzzy C-Means Clustering Technique

Nurul Illiyyana Emira, Jusoh (2015) EMG Classification Based On Features Reduction Using Fuzzy C-Means Clustering Technique. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 pages) EMG Classification Based On Features Reduction Using Fuzzy C-Means Clustering Technique 24 Pages.pdf - Submitted Version Download (376kB)

Abstract

This paper illustrates the features reduction technique for Electromyography (EMG) classification by using Fuzzy C-Means Clustering (FCM) technique. There are two types of EMG study which are diagnosis EMG and Kinesiological EMG. The diagnosis EMG is study about the characteristics of the motor unit action potential for duration and amplitude while for the Kinesiological EMG is about the movement analysis of the muscle activity. This work will focus more on Kinesiological EMG that used two types of electrode which are surface electrode and fine wire. The EMG signal is a measure of electrical current during the contraction of the muscle. EMG classification is not an easy task due to the signal contains a lot of uncertainties that leads to a high dimensional feature vector. The objective of this work is to extract the time domain features from the EMG signal and to perform the features reduction technique before classifying the EMG signal based on different pattern using FCM. In this work, five subjects that use right hand as dominant and without previous illness record are selected. Then, the subjects will be asked to perform 5 different patterns which are lateral, tripod, tip, power and extension. The EMG signal is collected at the forearm muscle. It is expected that FCM could perform EMG feature reduction to classify the upper limb muscle based on the different pattern.

Actions (login required)

View Item