3D Face Reconstruction Based On The Pose Estimation And A Single Image

3D face reconstruction technology is a key problem in the field of computer vision and computer graphics. 3D face reconstruction based on images is a mainstream technology.The technology based on multiple images has complex matching and fusion compared with that based on a single image, which has been a hot topic in recent years.But the modeling effectiveness and efficiency of current 3D face reconstruction based on a single image still can not meet the application requirements. This paper implements a 3D face reconstruction system based on pose estimation and single image, innovative work as follows:We propose a deformation algorithm for 3D face model from a single image using pose estimation.The method utilizes the attitude parameters to obtain depth information closed to the real face model, and make global and local fine-tune adjustment to an intensive 3D generic face mesh model.Experimental results show that the method can get a more realistic facial geometry,and can make a better balance between modeling results and modeling efficiency.We propose a texture information completion algorithm by use of mesh parameterization and face symmetry. The algorithm is for non-frontal face image texture missing problem, the key of face texture mapping is to establish correspondence between the 3D face model and the 2D face image, which combined with model parameterization method and facial symmetry information for the whole texture. Experiments show that this method can improve the effectiveness of face texture mapping in local area, and can solve the problem of texture information partial stretching deformation.