Abstract With the increase of subdivision depths, some problems of common data structures for representing the subdivision surfaces appear, such as excessive computer memory consumption and low efficiency of the data query which restrict the popularization and application of subdivision surfaces in more fields. By utilizing the topological characteristics of the subdivision surface, a two-layer data structure named CELL is presented in order to better realize the piecewise representation of trilateral/quadrilateral subdivision surfaces. The inner structure of CELL represents the subdivision surface patch by using arrays, and the outer structure of CELL represents the topological relations between the subdivision surface patches. Based on Catmull-Clark subdivision scheme, the structural compositions of CELL and the realization mechanism of the subdivision algorithm are proposed. Additionally, sharp and semi-sharp features are constructed, and a primary study on amalgamation of the image/Z-map model and subdivision surface is presented. The results of the experimental and theoretical analysis show the superior performance of CELL with relation to computer memory consumption, data query, subdivision surface computation and algorithm development.
Read full abstract