Toward robotic x-ray vision: new directions for computed tomography

Abstract

With the advent of industrial computed tomography (CT or reconstruction from projections), which is of much wider scope than medical CT, the basic mathematical and physical assumptions need to be reconsidered. This paper considers the ramifications of the indeterminacy theorem and the underdetermined nature of the reconstruction equations. A search for truth rather than honesty in their solution is suggested. The use of steered microbeams, CT with few photons, multimedia CT, and the CT of soft (deformable) objects are discussed. The latter subject suggests that CT should become amalgamated with finite element analysis and computer-aided design. The computational load of 3-D robotic x-ray vision may require fifth-generation computers.