3D object representation using cellular structured space modeling

Abstract

Modeling methods for 3D models in computer graphics usually start from geometry level, such methods can generate result displaying well on screen with no dubiousness, but these methods lack essential properties such as shape equivalence indicated by using topological properties of the models. Unfortunately, even though these topological properties are concerned, the efficiency for considering shape equivalence is quite low. In this research, we propose a valid shape modeling using cellular structured space, an abstraction level between topological and geometry level, for representing 3D objects in computer graphics. This proposed modeling method starts from creating set of zero dimensional cells or points and inductively compose a finite structure in each dimension to compose higher dimensional skeleton via attaching map until the maximum dimension of the models is three. The structure of this model is described by cellular structured table which canbe conversed to graph structure called CSSGraph. In experiment, 3D objects are constructed by using the proposed method and the result shows that the structure of the cellular structured space model is less complex than that of general geometrical model. Moreover, the similarity result shows that using the proposed model can categorize objects with shape equivalence more efficient than using only geometrical or topological data of the model