Image Reconstruction Techniques and Measure of Quality: Classical vs.Modern Approaches

Abstract

Mathematical methods are of central importance in the new technologies of image reconstruction. Some of the most important procedures are classified as Back-projection, Filtered Back-projection and iterative reconstruction techniques. Back- projection played an important historical role but is no longer used because of sizable artifacts. Analytical methods like Filtered Back projection excel in speed and accuracy when a large number ofprojections are available. These are extensively used in x-ray imaging. Algebraic Reconstruction Technique(ART) is more attractive when the number of views is limited and when noise is significant. For these reasons, iterative methods are widely used in imaging. Two slight variants of ART are SIRT (Simultaneous Iterative reconstruction Techniques) and SART (Simultaneous ART).A modern method of image reconstruction technique is Fast Slant Slack(FSS). This method is rapidly computable, algebraically exact, geometrically faithful and invertible. A new software known as beamlab is used for FSS image reconstruction. All these reconstruction techniques are explored in this work. Also, various tasks are performed to measure the immunity to noise and quality of the images using PSNR, MSE and Universal Image Quality Index.