Skeletonization using thinning method for human motion system

Abstract

Natural interaction is a form of human interaction with computers that shows human behavior towards computers. Kinematic presents the synthesis of three-dimensional computer graphics (3D) into the real world. Microsoft Kinect 3D sensor can help develop research in the field of kinematics by using RGB and depth images that can be used to improve the results of previous studies. Problems encountered in detecting kinematic is to determine the point features that will be used. The problems that exist in previous studies can be reduced by using the depth image generated by the Kinect. Depth image produces a simple shape that is used to simplify and accelerate the detection skeleton that can be used as features in the kinematic. Our method combines the advantages of the two methods, dilatation and star skeleton. Experiments show that our method efficiently and fast to extract the skeleton of a depth image that can be used as the kinematic features.