A hand shape recognition and its application to a GUI support system for a sight handicapped person

This paper presents the algorithms for hand-shape recognition and applies it to a GUI support system for a sight handicapped person. We propose two algorithms. The first algorithm is to recognize the angle of a forefinger only; and the second one is to recognize fourteen hand-shapes by using the method of skeletonization of a binary image. The former algorithm implements high speed processing to recognize hand-shapes in real time. As an application of this algorithm, we implemented a system in which the angle of a forefinger controls a reading point in a text on a screen in real time, and a text is read out with a sound synthesis software. Further, we propose an algorithm to recognize 14 types of hand-shapes to control further shapes in practice by using a binary image of the hand-shape.