Abstract

The bimodal optical computer (BOC) is a hybrid optoelectronic system that can solve a system of linear equations fast and accurately. The speed advantage of the BOC over the conventional digital computer is clear, especially when the system of linear equations is large or ill-conditioned. Applications for the BOC are numerous. In this paper, however, I choose the reconstruction of a two-dimensional image for computer tomography (CT). The CT image reconstruction problem, using the algebraic reconstruction technique (ART), can be written as a system of linear equations Ax = p, where A is an mn matrix that represents the intersection area of the object that needs to be scanned and the projection beam, and m is the number of projections. The object is divided into n nonoverlapping areas, x is an nl vector that represents the unknown density of the object, and p is an ml vector that represents the nonoverlapping projections. The solution of such a problem is given by x = (ATA)-1ATp. I have performed a computer simulation for this problem, and the results show that the BOC is able to solve such problem with great accuracy.

© 1990 Optical Society of America