부산대학교 도서관

This paper presents an efficient and stable vector-form finite element discrete analysis grid method, which can directly input the grid management with external characteristics for complex three-dimensional graphics. The experimental results show that the proposed deep learning algorithm can greatly improve the quality of mesh generation while maintaining the external characteristics of the shape and size of natural language. It reduces the quality of the area where the extra target surface is located and increases the movement speed of the model. The problem that the edge of the mesh extra target surface is turned over is solved, so that the topological relation of the mesh extra target surface layer by layer is more regular. The mesh is thus greatly improved.