Hand Waving Recognition System using Skin Detection Method by Near-IR Light

Mariko Takeuchia, Kunihito Katob and Kazuhiko Yamamotoc

Faculty of Engineering, Gifu University 1-1 Yanagido, Gifu, Gifu, 501-1193 Japan.


In this paper, we propose a recognition method of a hand waving gesture as a communication tool for a computer and a robot. We introduced the recognition of gestures that mean directionality such as “Come here” gesture and “Go away” gesture. These direct gestures are similar movement of a hand waving. First, we analyzed the movement of these gestures in order to find the feature difference between “Come here” gesture and “Go away” gesture. The results are the following:(1)Both gestures are cyclic movement of a hand waving. (2)Individuality exists in these gestures. (3)The different feature was obtained the time series variation of the wrist angle. The features are indicated by the changing model of elliptical feature of a hand shape. Then, we distinguished the two direct gestures by frequency analysis using obtained the movement feature and represent the recognition method is simple enough to recognize the direct gesture. Because the method is so simple, it is suitable for the gesture recognition that is required real time. We constructed the intuitive machine interaction system to control an event like operation of screen for recognition of the directionality of a hand waving gesture. For the system implementation, we used a skin material detection method by Near-IR light which is robust to background and individual difference of skin color information. The method is simple to extract the skin region by focusing on the unique reflection characteristic in materials of skin.

Keywords: Hand waving, Direct gesture, Frequency analysis, Skin detection method, Intuitive interface.

     Back to TOC