ウェアラブルディスプレイのための手首凹凸計測による手指の動的運動認識

抄録

To realize operation of a wearable display using hand motion, we have developed a dynamic-hand-motion recognition system using a wrist wearable device which measures wrist contour. In this system, time series data of wrist contour are visualized as a gray-scale image, and hand motion is recognized based on an image processing approach. The experiment in this paper examines the robustness of the system to the difference of hand motion time. Experiment results demonstrates that the system can recognize hand motion in a high accuracy (more than 90%) in case where the difference of hand motion time is not so large between training data and test data.