ABSTRACT
CT image reconstruction is a very interesting and challenging topic and an active research area. Novel algorithms are being continually developed. In this chapter, we first briefly review the history of reconstruction algorithms that can be traced back to as early as 1917, when J. Radon, an Austrian mathematician, first presented a mathematics solution for reconstruction of a function from these line integrals. However, his work did not attract much
attention, and no progress was made until the late 1950s, when the development of CT scanners gradually gained more attention in the medical community. Allan M. Cormack (1963) solved the problem of how to reconstruct images by using a finite number of projections, an important contribution. In the same year, William H. Oldendorf (1963) developed a direct backprojection method. Later, the idea of filtered backprojection was first proposed by Bracewell and Riddle (1967), probably the most influential development in this area. Gordon et al. (1970) proposed the algebraic reconstruction technique (ART), which may produce a good reconstruction when projections are not uniformly distributed or limited. During this period, a breakthrough was made by Godfrey N. Hounsfield at the Central Research Laboratory of EMI, Ltd., in England. During 1968-1972, he built the first CT scanner and obtained the first image of a patient’s head using an algebraic algorithm. For their pioneer work, Cormack and Hounsfield shared the Nobel Prize in Physiology or Medicine in 1979.
