ABSTRACT
This chapter describes few methods of averaging for image reconstruction. The main result with a statistical model of the image allows to achieve a high-quality image reconstruction through the process of iterations and renewal of statistics after each iteration. This approach does not require any transformation and its effective realization can be used for reconstructing the image in discrete form with any size. This image needs to be processed in order to filter such a noise, as is done, for instance, in the known method of filtered backprojection (BP). The chapter discusses the method of summation of projection data by the rays for the set of projections, which is described in the tensor representation of the image. The number of projections, or direction images, for image reconstruction by BP is considered to be 180, assuming the uniform change of the angle of projections by 1 degree.
