2D and 3D Super-resolution: A Bayesian approach

Abstract

Super-Resolution (SR) image reconstruction is the area of research and development which produces one or a set of high-resolution (HR) images from one or a set of low-resolution (LR) frames. Nowadays, there is a great number of methods of SR either for fixed or video sequence and for gray, color or hyperspectral images. There is, however, an area where we need still great effort is the 3D data, particularly in medical and biological or in non destructive testing (NDT) imaging systems where very often we may obtain low resolution 3D volumes reconstructed in different contexts (PET, Microscopy, molecular imaging, SAR imaging, holography and 3D TV, etc.) and we would like to reconstruct a HR 3D volume from a few LR ones. In this paper, first we report a few extensions to a recently developed Bayesian SR methods and then propose to extend it to 3D case.