3D Facial Model Reconstruction and Expression Generation Based on Dirichlet free-form Deformations Algorithm

With the development of computer technology, people have put forward higher requirements for online communication and other issues. Therefore, 3D face modeling and expression generation has become a hot spot in current research. How to generate face models with expressions mainly has the following problems: how to use face pictures to generate facial models and how to generate personalized expressions for different faces.To solve the above problems, this paper is based on the improved free-form deformation algorithm DFFD (Dirichlet free-form deformations), which realizes the reconstruction and expression generation of the face model using a single picture. Firstly, the MediaPipe framework is used to recognize the 3D coordinates of the facial feature points in the face image, and the facial model texture is realized based on the coordinate data, and then the improved DFFD algorithm is used to realize the expression generation. The experimental results verify the feasibility of the proposed method in 3D facial expression generation.