Constrained TV-minimization reconstruction from exterior CT data

Exterior data in CT refer to the data acquired from detector bins on one side of the detector center. It corresponds to, in the image space, an annulus-shaped scanning FOV. Exterior scanning geometry is of practical interest to obtaining the tomographic information of the peripheral region, when the center of the object can and/or should be spared from X-ray irradiation. In this work, we investigated the applicability of optimization-based reconstruction from exterior CT data. Data sets containing different numbers of detector bins from the edge, simulating different degrees of data exteriority, were extracted from the full, non-exterior data collected by scanning a hollow object with a laboratory CT system. The reconstruction results were characterized against image results using the conventional FBP algorithm. Quantitative assessments were carried out by calculating the UQIs and MSEs for all exterior reconstruction images with respect to reference images reconstructed from full data using the FBP algorithm. Both visual comparison and quantitative assessments suggest that optimization-based reconstruction is potentially a more viable solution for the exterior CT data than FBP reconstruction.