Hand gesture mining

Abstract

Hand gesture recognition and synthesis are both important for more natural and immersive Human Computer Interaction, many approaches to these tasks involve a large hand motion search space. Human hand motion is highly articulate and highly constrained. Although some constrains can be obtained by hand anatomy and biomechanics, it is an approximation of natural finger motion, in addition, these exist many constrains, which cannot be explicitly represented. In this paper, we presents a data mining approach to discovery the constraints of human hand motion, the result of the data mining is presented to display by VRML. Our presented approach is helpful for analyzing hand motion and hand motion capturing.