Ensemble registration: Incorporating structural information into groupwise registration

Abstract

We investigate incorporating structural information from segmentation into a groupwise registration framework. The aim is to augment conventional intensity-based registration, by including explicit structural information in the registration process. Our method uses various types of structural information, derived from the original intensity images. For the case of MR brain images, we augment each intensity image with its own set of tissue fraction images, plus intensity gradient images, which form an image ensemble for each example. We then perform groupwise registration using these ensembles of images. The method is applied to four different real-world datasets, for which ground-truth annotation is available. Various configurations of the ensemble are tested, and are also compared with a previously published method (which was only applied to the easier dataset), which used tissuefraction images to aid registration. It is shown that the method can give a greater than 25% improvement on the three difficult datasets, when compared to using intensity-based registration alone. On the easier dataset, it improves upon intensitybased registration, and achieves results comparable with the previously published method.