Abstract—The objective of this work is to develop a real-time gesture prediction system for navigation in a Virtual Reality Environment. Though earlier work presents situations where the level of activity is high, this research work refers to slight gestures, where the distinction is low. The paper further discusses the use of several machine learning methods to solve this prediction problem, including Support Vector Machines, Random Forests, and Artificial Neural Networks. After considering performance variation with respect to different test configurations, a conclusion is drawn on which configuration is to be used in the prediction engine. The final system was tested on a significant user base with results indicating sufficient accuracy and performance for real-time use.
Index Terms—Gesture input, virtual reality, supervised learning.
V. Dissanayake, S. Herath, S. Rasnayaka, S. Seneviratne, R. Vidanaarachchi, and C. Gamage are with the Department of Computer Science & Engineering, University of Moratuwa, Moratuwa, SriLanka. (e-mail: email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org).
Cite: Robert A. Sowah, Moses A. Agebure, Godfrey A. Mills, Koudjo M. Koumadi, and Seth Y. Fiawoo, "New Cluster Undersampling Technique for Class Imbalance Learning," International Journal of Machine Learning and Computing vol. 6, no. 3, pp. 205-214, 2016.