An Accelerometer-Based Approach to Evaluate 3D Unistroke Gestures
Tahir Madani1,2,*, Muhammad Tahir3, Sheikh Ziauddin1, Syed Raza1, Mirza Ahmed1, Majid Khan1, and Mahmood Ashraf 4,5
1Department of Computer Science, COMSATS Institute of Information Technology, Pakistan
2Department of Information Technology, University Technology Petronas, Malaysian
3Faculty of Computing and Information Technology, King Abdulaziz University, Kingdom of Saudi Arabia
4Department of Computer Science, Science and Technology, Pakistan
5Department of Software Engineering, Universiti Teknologi, Malaysia
Abstract: This paper presents an evaluation of Three Dimensional (3D) unistroke human arm gestures. Our scheme employs an accelerometer-based approach by using Nintendo™ Wiimote as a gesture device. The system takes acceleration signals from Wiimote in order to classify different gestures. It deals with numeric gestures, i.e., digits from 0 to 9 and simple mathematical operator gestures for addition, subtraction, multiplication and division. Two techniques, Dynamic Time Warping (DTW) and 2D trajectories are used to recognize and classify gestures. Successful recognition rates indicate that performing 3D gestures using accelerometer-based devices is intuitive and provides an effective means of interaction with computers.
Keywords: Human computer interaction, 3D gestures, accelerometer, 3D calculator, dynamic time warping, 2D trajectories, wiimote.
Received May 29, 2012; accepted September 26, 2013