A minimum cross-entropy type algorithm including a model of the object
Christine Mello, Ricardo M. Campello de Souza · 1992
Summary form only given, as follows. A minimum cross-entropy type algorithm including the a priori knowledge of the project-in the form of a model-is proposed for application in computerized tomography. This algorithm deals with incomplete projection or limited angle data. To obtain the MIDINMOD (minimum divergence with a normalized model) the cross-entropy functional is minimized using as a priori distribution a normalized model of the object. The performance has been compared with that of the MENT and EXTEND MENT algorithms.>