Methods of recognition were compared in this paper for characteristic signals of hand waving gestures. Confused motions between “beckoning” and “waving away” were distinguished by characteristic signal mixing with a gravity motion and an area variation in the image of a moving hand. Their dynamic gestures are recognized by DP matching, wave difference, and other methods. Results were as follows: (1) taking the difference of a hand waving trajectory into account, the recognition ratio rose to more than 92.3%, (2) but by using categorized vectors to search for the same vectors in an input waving motion, it was made clear that the missed-recognition ratio between “beckoning” and “waving away” was suppressed.

1. [1] N. Kashiwagi, K. Nishihara and N. Kawarazaki: “Manipulator Work System using Vision Sensors,” Proc. Robotics & Mecatronics, JSME, [No.98-4] 2Cll pp.1-7, 1998-6.
2. [2] Y. Kuniyoshi, M. Inaba and H. Inoue, “Seeing, Understanding and Doing Human Task” : Proc. of the IEEE Int. Conf. on Robotics and Automation, pp.2-9,1992.
3. [3] T. Yokokawa, et al, “Real Time Estimation Pointed Direction in 3D World using Visual Sensor,” Proc. Robotics & Mecatronics, JSME, [No.97-22] Vol.A, 1997.
4. [4] K. Imagawa, et al, “Real Time Tracking of Human Hands from a Sign-Language Image Sequence in Consideration of Disappearance by Skin Regions,” Trans. Inst. of Elec. Inform. and Comm. Eng. D-II Vol.J81-D-II No.8 pp.1787-1795, 1998-8.
5. [5] K. Nishihara, H. Kitajima and N. Kawarazaki, “A Feature Extraction and Recognition of Hand Waving Gestures by a Visual Camera,” Trans. JSME (in Japanese), C-67, 600, pp.215-221, 2001.