Gesture recognition by a remote computer controller

Бесплатный доступ

The remote controller is based on MEMS-accelerometer tracks three dimensional movements of an operator hand. The controller driver generates a signal that controls the mouse cursor. It is shown that an intelligent driver uses the control signal repeatedly to monitor the operator actions and builds a portrait of operator arbitrary actions as a sequence of his hand gestures. For this purpose, pattern discovery algorithms are used for processing control signal records.

Remote manipulator, accelerometer, control signal, prony method, mueen-keogh motif discovery algorithm, алгоритм mueen-keogh motif discovery, pattern discovery, cuda, dynamic time warping

Короткий адрес: https://sciup.org/142185847

IDR: 142185847

Статья научная