Vision-based Hand Gesture Recognition for Mobile Service Robot Control

Abstract

The vision-based hand gesture is an important problem in the field of humancomputer interaction since hand motion and gesture could potentially interact with computers in more natural ways. This paper is focused on control of mobile service robot based on visionbased hand gesture recognition. The database of five hand gesture commands is created and the cross-correlation find the similarity between the captured hand image. This approach contains image pre-processing, segmentation of the region, feature extraction, classification using crosscorrelation and gesture recognition. In this method, the user can control the robot by using the gesture of the hand, thereby interacting with the robot system. The commands are generated according to the hand gesture for controlling the robot direction. Experimental results show that this method can get a high recognition rate of hand gesture.