Finger motion gesture recognition system as a new human computer interface

Abstract

A gesture-based interface, in which the user specifies commands by simple drawings, offers an alternative to traditional keyboard, mouse, menu, and direct manipulation interface. A single intuitive gesture can simultaneously specify objects, an operation, and additional parameters, making gestures more powerful than the "clicks" and "drags" of the traditional direct manipulation interfaces. Gesture-based interfaces researched so far have needed a mouse or stylus pen as a gesture input device. However, the author strongly believe that the concept of making gestures with the human finger appeals more not only to experienced users but also to the novice users who are usually not accustomed to using computer devices. This dissertation describes the finger mouse system(FMS) and finger gesture recognition system(FGRS). FMS emulates the pointing and selecting functions of a mouse by capturing the image of a hand and analyzing its motion and shape. FGRS provides the user with the capabilities of command input and direct manipulation by recognizing the gesture which is a sequence of points input by FMS. The author utilized syntactic pattern recognition technique for implementing FGRS and designed it for helping programmers build gesture-based applications easily. In addition, the author introduce an example drawing program in order to demonstrate the practicality of the author‘s approach.