Similarity Metrics

As highlighted in the previous chapters, the goal of image registration is to find a geometric transformation (rigid or nonrigid) that aligns two given images. When two images are deemed “registered,” this indicates that they are best matched or most similar compared with the initial relationship between the images. With this concept in mind, similarity metrics are arguably the most critical element of a registration problem. The metric defines what the goal of the registration process is, and it measures how well one image is matched to another after the transformation has been applied. The aim of this chapter is to describe some of the basic similarity measures that are used in the process of image registration.