Tutorial 9: Introduction to Image Registration
Presented by
William Wells, Massachusetts Institute of Technology and Lilla Zollei, Massachusetts General Hospital
Abstract
We plan to provide an introduction to the problem and solution of image registration in a lecture-style tutorial. We will cover three main topics: transformation spaces, objective functions and optimization methods. For each of these, we will explain how they are related to the correspondence problem and how their choice influences the efficiency and/or performance of registration algorithms.
In the last hour of the tutorial we will discuss a bit more advanced material. We will go over the historical development and importance of the mutual information registration function and also provide a theoretical framework which helps to situate it with respect to other popular objective functions currently used in the community.
Lecture notes will be distributed suring the tutorial session. In addition, supplementary material and a limited set of relevant hands on registration excercises can be found on the website of the Open Courseware class, HST 582, Biomedical Signal and Image Processing.
