Research on Realistic 3D Face Rapid Modeling Based on Animation Symbols

3D face modeling is a complex subject, especially in the animation and game industries. It is of great significance to quickly generate realistic 3D face models for improving the visual effect of works. Realistic 3D face modeling often involves complex mathematical models and algorithms, which leads to a long time-consuming modeling process, high matching accuracy and high modeling average error. In order to solve the above problems, this paper proposes a research on realistic 3D face rapid modeling based on animation symbols. The feature set of 3D animation face is constructed, the one-to-one mapping relationship between 3D mesh and 2D plane is established, and the method of triangular deformation and bilinear difference is introduced to realize the feature point mapping between covering texture and neutral texture. In order to get a realistic 3D face model, we use the fusion shape model of interpolation to generate the key facial action target mesh, use the weighted DFFD deformation algorithm, assign weight to the Sibson coordinate components, and combine the local organ data to adjust and correct the 3D contour face. The experimental results show that the proposed method can accurately calibrate the feature points of 3D animation faces, the matching accuracy is the highest 98.39%, the modeling time is only 0.5[Formula: see text]s, and the average modeling error is the lowest about 0.32%.