Skip to Main Content
We have developed two new "meta-algorithms" for computed tomography that give significantly improved images through deconvolution of the two-dimensional point spread function of standard, quasi-linear algorithms. In geometric deconvolution the projections of the point spread function provide the basis for a set of one-dimensional deconvolutions. In two-dimensional Wiener deconvolution, the two-dimensional point spread function is deconvoluted directly. The criticism that there is no data available for these deconvolutions is met here by showing that the "missing data" is partly provided by incorporation of a priori information, as is the practice in other superresolution work.