An interpolation algorithm fitted for dynamic 3D face reconstruction

In order to solve the problem of low recognition accuracy in later period which is caused by the too few extracted parameters in the 3D face recognition, and the incapable formation of completed point cloud structure. An automatic iterative interpolation algorithm is proposed. The new and more accurate 3D face data points are obtained by automatic iteration. This algorithm can be used to restore the data point cloud information of 3D facial feature in 2D images by means of facial three-legged structure formed by 3D face and automatic interpolation. Thus, it can realize to shape the 3D facial dynamic model which can be recognized and has high saturability. Experimental results show that the interpolation algorithm can achieve the complete the construction of facial feature based on the facial feature after 3D dynamic reconstruction, and the validity is higher.