Overview
The
registration of volumetric data from sources as MR, CT or PET, is of importance
for surgical planning, diagnosis, clinical track of events, evaluation of
radiotherapeutical procedures and medical research. Alignment can be achieved in
many ways, one of which is by maximizing a measure of similarity between images,
such as the mutual information similarity. This technique requires no a
priori model of the relationship between modalities, it only assumes that one
volume provides the most information about the other one when they are correctly
registered.
This project
presents a wide overview of existing medical image registration methods.
Registration methods are classified according to the different aspects of
mutual-information based registration. Three state-of-the-art
registration software packages, available for public use, are explored. The
software was compared in terms of registration accuracy (using consistency
measurements), visual inspections, and by applying, for the first time, clinical
tests on Multiple Sclerosis (MS) patients.
Our thanks to
Dr. Anat Achiron, Multiple Sclerosis center, Sheba Medical center, Tel-Hashomer,
Israel.
