Growing Grid-Evolutionary Algorithm for Surface Reconstruction

This work primarily aims at introducing an algorithm for surface construction in conjunction with hybrid Growing Grid network and Evolutionary Algorithm, called Growing Grid-Evolutionary network. The process of surface construction primarily consists of two main steps namely: parameterization and surface fitting. The application of growing grid network is implemented at the parameterization phase; meanwhile the evolutionary algorithm has been used to optimally fit the surfaces through the Non Uniform Relational B-Spline (NURBS) method. Various graphical data are used in the experiment including the free-form objects, parabola, and mask. In order to validate the proposed algorithm, we conduct an error analysis for each step of parameterization and surface fitting by comparing the surface images generated with the original surfaces. Experimental results show that the proposed growing grid-evolutionary network has successfully generated surfaces that resemble the original surfaces and enhance its performance.