Hybrid Computing Algorithm in
Representing Solid Model

Muhammad Matondang and Habibollah Haron
Department of Modeling and Industrial Computing, Universiti Teknologi Malaysia, Malaysia

Abstract: This paper presents an algorithm, which is a hybrid-computing algorithm in representing solid model. The proposed algorithm ‎contains two steps namely reconstruction and representation. In the reconstruction step, neural network with back propagation ‎has been applied to derive the depth values of solid model that was represented by the given two-Dimensional (2D) line ‎drawing. And then in the representation step, once the depth value was derived, the mathematical modeling was used to ‎generate the mathematical models to represent the reconstructed solid model. The algorithm has been tested on a cube. ‎Totally, there are eighty-three cubes has been used on the development of neural network model and six mathematical ‎equations yielded to represent each one cube. The proposed algorithm successfully takes the advantages of neural network and ‎mathematical modeling in representing solid model. Comparison analysis conducted between the algorithm and skewed ‎symmetry model shows that the algorithm has more advantages in term of the ease of the uses and in simplifying the use of ‎mathematical modeling in representing solid model.‎

Keywords: Solid model, reconstruction, representation, neural network, mathematical modeling, and hybrid computing.‎