A human-machine interface for manipulating a virtual 3-D object using natural hand gestures

摘要

In this paper, we propose a new human-machine interface for manipulating 3-D contents in a display. Motions of a hand or both hands of a user are taken images by a simple USB camera. The motions are recognized by a PC and transformed into corresponding instructions such as translation, rotation and scaling. In addition, unintentional movement and unconscious movement are eliminated in the method and smoother operation is realized. The result of the manipulation is displayed in the screen. Experiments were performed using Google Earth and 3-D CG contents, and the proposed system successfully manipulated its rotation and scaling by some hand motions of the user.