Block Matching Monomodal Image Registration using Robust Similarity Measure and a combination of optimization and interpolation

In this paper we present a Block Matching approach to registration of medical 2D images IRM/IRM. The registered images are assumed to be rigidly aligned before starting this procedure. The sum of absolute differences (SAD), sum of squared differences (SSD), mutual information (MI) and correlation coefficient (CC) are used as measures of similarity to determine the similarity between images as well as to evaluate the degree of robustness of registration. In order to provide the best value of a measure of similarity, process of optimization and interpolation are introduced. The duration of the algorithm’s execution is dependent on the block’s size. The objective of this article is to choose the best suitable measure of similarity and to test the effect of subdivision of blocks on the duration of execution which is for the benefit of medicine. This approach was tested and leads to interesting results.