Gestures in Machine Interaction

Vnencumbered-gesture-interaction (VGI) describes the use of unrestricted gestures in machine interaction. The development of such technology will enable users to interact with machines and virtual environments by performing actions like grasping, pinching or waving without the need of peripherals. Advances in image-processing and pattern recognition make such interaction viable and in some applications more practical than current modes of keyboard, mouse and touch-screen interaction provide. VGI is emerging as a popular topic amongst Human-Computer Interaction (HCI), Computer-vision and gesture research; and is developing into a topic with potential to significantly impact the future of computer-interaction, robot-control and gaming. This thesis investigates whether an ergonomic model of VGI can be developed and implemented on consumer devices by considering some of the barriers currently preventing such a model of VGI from being widely adopted. This research aims to address the development of freehand gesture interfaces and accompanying syntax. Without the detailed consideration of the evolution of this field the development of un-ergonomic, inefficient interfaces capable of placing undue strain on interface users becomes more likely. In the course of this thesis some novel design and methodological assertions are made. The Gesture in Machine Interaction (GiMI) syntax model and the Gesture-Face Layer (GFL), developed in the course of this research, have been designed to facilitate ergonomic gesture interaction. The GiMI is an interface syntax model designed to enable cursor control, browser navigation commands and steering control for remote robots or vehicles. Through applying state-of-the-art image processing that facilitates three-dimensional (3D) recognition of human action, this research investigates how interface syntax can incorporate the broadest range of human actions. By advancing our understanding of ergonomic gesture syntax, this research aims to assist future developers evaluate the efficiency of gesture interfaces, lexicons and syntax.