Three-Dimensional Hand Tracking Based on Microstructure of Hand Motion

At present, hand tracking based on particle filter obtains current 3D hand gesture mainly by two elements, that is, by current frame of gesture images and by status information of the previous 3D hand gesture. Compared with current algorithms, the proposed algorithm focuses on behavior analysis and process modeling. Focusing on mathematical linkages establishment between microstructure model and the current frame gesture, the status of current 3D hand gesture is obtained not only on the basis of local frame information, but gesture model of overall movement process. After that, a unified and efficient data structure is presented, upon which the particle number can be reduced. First, hand motion law in specific circumstance is obtained by experiments based on data glove in the three-dimensional virtual platform. Second, the law is summarized as the microstructure of hand motion. Hand motion model is put forward and applied to forecast period of the particle filter algorithm, upon which hand tracking can be guided by the summarized microstructure. At last, hand tracking algorithm is carried out by a camera-oriented virtual three-dimensional interactive platform. The experimental result shows that, compared with Gaussian Process Annealed Particle Filter (GPAPF) gesture tracking algorithm, the time cost of the proposed algorithm was nearly 29% lower, while the accuracy 26% higher.