Identification of hand gestures using pattern recognition of electromyography signals acquired with MyoArmband

Abstract

The most used way to record hand gestures' information is through the Electromyography (EMG) technique. However, the research in the area is still fragmented. This study aimed at reproducing the high classification performance of hand gestures using EMG data reported in the literature, using a MyoArmband EMG equipment for data acquisition and an LDA classifier, and testing different features and feature selection techniques. The results showed that a performance of 75% is achievable with selected features.