2008 , Volume 13, № 6, p.121-133

Problems of computational tomography with a small number of views require application of algorithms utilizing a priori information regarding the solution. Gerchberg-Papoulis algorithm (G-P), based on the theorem of a central slice, is known to be one of the best iterative methods of a few-view tomography for a parallel geometry of scanning. In a fan-beam tomography case such algorithm has not been yet investigated, because of there is no an analog of the central slice theorem for this case. In the present contribution, a recently developed generalization of the central slice theorem for fan-beam geometry is described. A new iterative G-P algorithm is developed based on the generalization. Two versions of the numerical scheme are investigated. An influence of the superimposed random noise on accuracy of reconstructions is studied; regularization criteria are developed which allow suppression of noise in measurements.