Research On 3d Face Geometry Optimization Based On Artificial Intelligence And Computer Graphics

The establishment of high-quality 3D models is one of the critical issues in Computer Graphics research,similarly,the application of 3D face models has always been a significant topic of discussion in Computer Graphics and Artificial Intelligence research.In essence,this paper optimizes the geometry of a 3D face mesh model to improve the usability of the 3D face model data.Aiming at the hole problem of the 3D face model,taking the uniqueness of the biological characteristics of the human face into account,using the method of facial landmark detection in artificial intelligence,combining the 3D to 2D mapping relationship,and extracting the 3D face feature area mesh,this paper uses the face feature information to process the 3D face model,thereby improving the integrity and robustness of model repair.In this paper,Artificial Intelligence and Computer Graphics are combined to optimize the geometry of a 3D face mesh model.This research designs the 3D face mesh model for quality evaluation system framework,proposes the feature extraction and matching method of the 3D face model based on facial landmark detection and then merges the semantic information to realize the research on the hole repair of the 3D face mesh model based on the Poisson equation.The main research work of this paper is described as follows:(1)This paper combines Artificial Intelligence with Computer Graphics to propose a 3D face mesh feature extraction and matching method based on facial landmark detection.This method combines the facial feature recognition algorithm of the 2D face image with the mapping relationship of the 3D face model and realizes the semantic segmentation and extraction of the 3D face model features,which provides a basis for the next step of the 3D face mesh repair research.(2)This paper proposes a quality evaluation system based on a 3D face mesh model to detect the quality problems of 3D face models.This system came into existence to establish a complete 3D face model database in the future.First,the system evaluates the quality of the3 D face mesh model data and then classifies the obtained results to improve the availability of the 3D face model.(3)The feature enhancement of the 3D face mesh model based on the Poisson equation effectuates the hole repair.In this paper,the hole detection of the 3D face model is carried out,and then,the boundary of the area mesh is marked to determine the hole area of the model.According to the working principle of the Poisson equation and the application of the isosurface extraction technology,this research extracts the hidden layer weights in the face detection model convolutional neural network,calculates the eigenvectors of each region,and enhances the surface fitting of the Poisson equation by introducing the eigenvectors to realize the hole repair of the 3D face mesh model based on the Poisson equation and optimize the 3D face mesh model.