Inner distance based hand gesture recognition for devices control

摘要

Body language is a connection way between humans. Computer vision based system try to facilitate human connections using human gesture interpretations. This paper proposes a method based on hand gesture recognition for device control using inner distance feature. Our method consists of three primary parts. 1) Hand segmentation and state identification. 2) Hand tracking, and 3) gesture recognition. In tracking phase, we propose a new method based on hand center of gravity tracking and in recognition phase, we use correlation coefficient for state matching.