Hand gesture recognition system using depth data

Abstract

Ongoing efforts at our laboratory aim at establishing a gesture recognition system to control the movement of mobile robot The proposed work is recognizing several predefined gestures by using a sequence of depth images which obtain from Kinect We present a new method to recognize hand gesture including five main steps: gestures segmentation, feature extraction and matching, establish 3D coordinate, trajectory extraction and classification. Experimental results are shown that the system can be used to effectively control the movement of the mobile robot.