Texture Mapping And Reconstruction Of 3D Face Model Based On Feature Constraints

3D face reconstruction has always been an important research content in the fields of computer graphics and computer vision.It has a wide range of applications in film and animation,3D games,face recognition and so on.Compared with the 3D face reconstruction algorithm based on image sequence and video,the 3D face reconstruction method based on single face image has less user input,so it is a hot topic in the research of face 3D reconstruction.However,due to the lack of depth information in a single face image,how to reconstruct a realistic three-dimensional face in a relatively short time with a low computational complexity is still challenging.Aiming at this situation,this paper proposes a method based on feature constraint for 3D face model texture mapping and reconstruction.The main work contents are as follows:(1)The face image feature point automatic extraction and 3D thin plate spline function are used to complete the 3D shape reconstruction of the face,and the face model library is used to establish the general 3D face model and restore the depth information of the image feature points.Because the deformation of the face model will be unsmoothed by using the thin plate spline function directly,this paper proposes a method of smoothing the thin plate spline function to reconstruct the 3D face model,so that the 3D face model recovers the shape with good smoothness at the same time.(2)A parameterization method of conformal plane based on domain polygon is proposed,and the mapping relationship between 3D mesh model and 2D plane domain is established.The Poisson fusion method is used to seamlessly fuse with the standard skin color image to obtain the texture map,and the texture information of the 3D face model is restored by the least squares coarse alignment method.Aiming at the texture offset of the feature areas such as eyes and mouth after coarse alignment,this paper proposes a method of segmentation optimization for face feature region,and finally obtains the result of realistic texture mapping.(3)We compared the realism and time efficiency of face model with different reconstruction methods,and introduced the application of 3D face reconstruction in 3D fitting.The experimental results show that the proposed method of the 3D face model texture mapping and reconstruction based on feature constraints can reconstruct the realistic 3D face model in a short time.Finally,the main research contents of this paper are summarized,and the future work focus is pointed out.