Masters Thesis

Hand gesture recognition using unsupervised learning

Gesture recognition is one of the revolutionary technological advancements seen in human-computer interaction field today. The applications of gesture recognition are vast from sign language recognition, through prosthesis control, to control interface for virtual reality (VR) and augmented reality (AR) systems. Many of these applications use supervised learning-based machine learning algorithms to interpret data and make decisions. Unsupervised learning is still in the early development stage since it involves streaming in continuous unlabeled data sets and obtaining meaningful data sets out of billions of data remains a challenge. This project aims to develop and analyze k-means clustering - an unsupervised learning technique, for hand gesture recognition based on unlabeled electromyographic (EMG) data sets obtained from a commercial armband Myo developed by Thalmic Labs.

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.