Incorporation of a priori moment information into signal recovery and synthesis problems

AbstractThe problem addressed is that of synthesizing a function subject to a priori moment constraints. Each moment constraint requires that the function to be synthesized lies in a closed convex set. For r moment constraints, the function must lie at the intersection of r convex sets. The method of projections onto convex sets (POCS) is used to synthesize a function consistent with the moment constraints. The algorithm has application in medical imaging, especially in computer aided tomography