Unifying Models and Registration: A Framework for Model-based Registration and Non-rigid Registration Assessment

Statistical models of shape and appearance are widely used for the analysis of biomedical images. Two deficiencies of these models is that they require annotation across a large number of images in order to be built and, having built such models, it is then difficult to reason about their validity or assess their quality. Herein, a method is described which addresses both problems and establishes a unified solution. In order to construct the models rapidly, corresponding structures must be brought into the state of dense overlap. Image registration is the mechanism whereby a set of images can be analysed in a common frame of reference and models then derived from it. The thesis provides a solution to the recurring need to compare such models and extends the method as to provide an image registration assessment method, which does not require ground truth. The thesis also deals with a complementary case where images are registered by minimising the complexity of models. Overall, the proposed framework can be perceived as one which ucombines registration and modelling, taking advantage of the fact that they are innately the same. Registration provides correspondence across images and, given that correspondence, models of appearance can be built and registration