Image registration is an emerging topic in image processing with many applications in medical imaging, picture and movie processing. The classical problem of image registration is concerned with finding an appropriate transformation between two data sets. This fuzzy definition of registration requires a mathematical modeling and in particular a mathematical specification of the terms appropriate transformations and correlation between data sets. Depending on the type of application, typically Euler, rigid, plastic, elastic deformations are considered. The variety of similarity measures ranges from a simple Lp distance between the pixel values of the data to mutual information or entropy distances