ABSTRACT
Due to practical limitations in data acquisition, 3-D computed tomography systems must attempt to provide rapid reconstructions of acceptable quality from a limited number of views. The use of convolution backprojection (CBP) for image reconstruction from an inadequate number of projections, results in view aliasing artifacts. In this paper we investigate different post-processing methods of alleviating the effects of view aliasing artifacts. Two distinct methods and their variants are considered. The first, termed the intermediate view reprojection (IVR) method, involves estimating a set of intermediate views by reprojection, followed by a reconstruction using the augmented set of views. The second, termed the error-correction (EC) method, incorporates a correction on the initial reconstruction based on the projection-domain error. Suitable modifications and variants of the above methods are indicated. Of the methods discussed, the IVR method is simple, tends to reduce the effects of artifacts with less susceptibility to secondary effects, and is applicable to region-of-interest reconstructions.