Finding a set of optimal viewing for 3D surface reconstruction from 2D X-ray images using genetic algorithm

Abstract

In 3D reconstruction, the efficiency and the accuracy of the reconstruction depend on the distribution of the reconstructed surface points or landmark points. This paper presents a new method, using multi-cellular genetic algorithm, to determine a set of any chosen number of projection angles that approximately provides the optimal distribution of the reconstructed surface points. This is to maximize the use of a small number of 2D X-ray to reconstruct surfaces and allow all important features to be reconstructed by using a minimum number of views. Some proof of concept computational experiments have been conducted to demonstrate the capability of the proposed method.