Preliminary study on the hand clapping action recognition based on the leap motion controller

Abstract

In this paper, we propose a method of recognizing human hand clapping action based on the Leap Motion Controller. Basic hand clapping is a kind of simple but somewhat fast periodic hand action. By the inevitable occlusion of two hands, fast speed of hand motion, and its neighboring noisy data, it is not easy to be recognized either by vision sensors or by sound sensors. Leap Motion Controller (LMC) is one of the most advanced and low-priced devices which can be used to detect the detailed motion of multiple moving objects in some restricted range. By the fast and accurate hand trajectory data from Leap Motion Controller, we can predict the relative position of two palms more accurately. And, by using the prior knowledge about basic hand clapping action sequence, LMCbased hand clapping recognizer has been developed. The average success ratio is 86.11% with the three different test users and 240 tests per each user. © 2014 CSREA Press.