Abstract
In many radiation imaging applications, such as X-ray computerized tomography or nuclear medicine, tomographic systems have been recently developped, which image an object from an angulary restricted number of projections. It is well known that in such a case, some Fourier components are missing {1}, and thus analytical reconstructions, such as Fourier synthesis can no longer be used. But, the algebraic methods allow to take advantage of "a priori" informations on the object, in order to recover the missing data. Many authors have proposed different reconstruction techniques, that combine limited projection data with a priori object information through iterative revisions in both image and transform spaces {2-3}. These methods in general consume much computer time.