Rigorous Error Bounds in Computerized Tomography

Practically speaking, how can one utilize x-ray projection data so that the reconstruction obtained is in fact an approximation to the attenuating object? An answer is given which depends on the use of special projeition angles but provides an upper bound in the L2 topology for the difference between the reconstruction and the x-ray attenuating object. The keys to this solution include two theorems and an important assumption: some a priori knowledge of the attenuating object. The article is somewhat of a review of [3], but focused on the development of the error bound.