PDF version of this entire document

Conclusions

A method for registering images in a groupwise fashion has been described, including the quality of their appearance model in the objective function. Not only does this enable an algorithm to reach good results from a groupwise perspective, but it also results in the automated construction of appearance model, whose quality is assessed.

A model-based approach to evaluating the results of NRR of a group of images has also been described. The most important advantage of the new method is that it does not require any ground truth, but depends solely on the registered images themselves.

The approach has been demonstrated by studying the effect of perturbing, progressively, the registration of an initially registered set of images, comparing the results with those obtained using a `gold standard' measure based on the overlap of ground-truth anatomical labels. It has been shown that the new method provides measures of registration accuracy that are monotonic functions of the known misregistration, and that one, Specificity, provides a more sensitive measure of misregistration than the approach based on ground truth.

The model-based approach requires a distance measure in image space, and it has also been demonstrated that the use of shuffle distance, rather than Euclidean distance, improves the sensitivity of the approach.

A set of experiments have further validated the approach, and illustrated its application, by performing a comparative evaluation of the results obtained using three different NRR algorithms, demonstrating the superiority of a fully-groupwise algorithm over a repeated pairwise approach.

It is important to emphasise that this approach is not restricted to evaluating model-based NRR algorithms, though the thesis presented results for one such method; the model-based measures of registration accuracy can be applied to any set of non-rigidly registered images, however they were obtained. The possibility of a bias in favour of model-based methods of registration has been discussed. It was concluded that there is no major problem, though it would be desirable to compare results obtained using a range of ground-truth-free methods of evaluation.