Computed Tomography (CT) has been one of the most exciting imaging modalities developed in the biomedical field. Although the problem of imaging and image reconstruction arises in many scientific fields, e.g. astronomy and non-destructive testing, its most prominent applications are found in the field of diagnostic medicine. Imaging from x-ray projections is the process of reconstructing an image of the two-dimensional distribution of x-ray attenuation coefficient from measurements of projection data which are line integrals along a finite number of lines of known orientations. There have been significant advancements in CT scanners, with regard to both physical components as well as reconstruction algorithms. Based on the well-tested Filtered Back Projection approach, a new CT imaging system is developed to enable the reconstruction of objects from multiple projection data obtained from a GE vertical-stationary x-ray imaging system. Although this system is not designed for CT purposes, in this thesis we present a low cost scheme for obtaining multiple view projection, from which to reconstruct the object. There are many artifacts introduced by the instrumentation of the system for CT imaging. These issues are investigated and algorithms are developed to compensate for these artifacts caused mainly by mechanical errors during rotation. Modifications in adaptive filtering algorithms are also incorporated for further improving the reconstruction results. Thereby, an accurate and low-cost CT imaging system is realized. -- Abstract.
Read
