Design and Development of Hand Gesture Recognition System

Abstract

In this paper a vision based hand gesture recognition system is developed. Here the computer perceives the human gestures and
responds in a pre-defined way. It uses the Kinect sensor built on a software technology developed internally by a subsidiary of Microsoft Games
Studios owned by Microsoft. The proposed algorithm consists of a sequence of steps which are further divided into sub-steps, each of which is
performed using different algorithms. The major steps include preprocessing, hand segmentation and feature extraction. In this work the
segmentation of the hand from the rest of the image is carried out. Also elimination of the background noise, calculation of centroid of the hand,
identifying the gesture and matching it to the stored templates is simulated. To capture the images concepts like contours, convex hull and
convexity defects are applied. The paper provides a prototype on how different gestures of the hand can help interact with the computer interface
easily.

Keywords: Contours; Hand Gesture; Image Processing; Kinect sensor; Preprocessing;