Fingertip positioning and tracking based on motion prediction with monocular camera

This paper presents a method of fingertip positioning and tracking. Image sequence is acquired from a monocular camera. There are three steps to eliminate the interference, which are differential processing, noise reduction and shadow removal. Then we take corner detection and elliptic curve fitting to determine the fingertip position. In fingertip tracking, we use an improved compressive tracking algorithm based on motion prediction, which makes it possible to remove the false tracking points. Simulation results show that the proposed algorithm can correctly locate and track the fingertip at moderate moving speed in the laboratory scenario.