A Unifying Framework for Reconstructive Imaging Methods

Reconstructive imaging methods, such as CT or MRI, produce their images in a two stage process: First, the object to be imaged is probed using some form of physical radiation, and then the image is computed, or reconstructed, from the outcome of the probing. The computation is based on a reconstruction algorithm which tries to invert a mathematical model of the probing. In this paper we inspect and study these building blocks of reconstructive imaging methods in more detail. Special attention is given to an unavoidable property of the reconstruction problem, known as ill-posedness. As a result, we obtain something like a unifying framework for reconstructive imaging methods that helps to better understand, design and estimate these methods. Finally, the framework and its applications are briefly illustrated by way of some examples.