| Computed tomography is a medical imaging method to provide cross sectional images from x-ray projections. The first mathematical formulation for reconstructing an object from multiple projections was developed by Radon in 1917, and the first clinical CT scanner was introduced in 1971. Since then, the priorities of CT scanner development were to achieve faster scan times, improved dose efficiency and improved image quality. Even though technological innovations have provided dramatic improvement during the last few decades, artifact free volumetric imaging in a single rotation is still a challenging issue.;To achieve these goals, a novel concept for a CT system, multi-source inverse-geometry CT (MS-IGCT) has been proposed. Instead of using a single x-ray source and a large detector array, the MS-IGCT system employs a distributed source array and a smaller detector array. Since the source and detector array have the same extent in the axial direction, acquired projection data satisfy the data sufficiency condition for an exact reconstruction, and therefore image artifacts inherent in 3rd cone beam CT system can be reduced. In addition, the smaller size of the detector array can reduce the detected scatter radiation, and the radiation dose to the patient can be optimized by adjusting the x-ray flux of each source independently.;This work consists of three parts. The first part presents a new normalization method for the MS-IGCT system. For raw data normalization, two approaches will be presented, and compared with the ideal case. The second part presents several reconstruction algorithms for the IGCT system. The main focus of the reconstruction algorithms is to improve the image quality (i.e., SNR and resolution), and several approaches are presented. The final part addresses the noise power spectrum (NPS) of divergent beam CT systems. Because of the non-stationary noise property of a divergent CT system, the NPS of a divergent CT shows the high frequency roll off, and the degree of the high frequency roll off shows different behaviors according to the size of reconstruction FOVs. In this part, an analytical expression of the NPS for a divergent CT system is derived and compared with computer simulations. |
