PDF version of this entire document
The results of assessing the generalisation and specificity for each of the three models is shown in Fig. 15. This shows that the full groupwise method is better than the partial method (without shape constraints), which in turn is better than a simple pairwise approach. The evaluation technique allows us to compare different algorithms and make quantitative judgements on the effect of different approaches.
The results of the experiment to test the effect of increasing mis-registration were shown in Fig. 11 and Fig. 12. These demonstrates that, for all sizes of shuffle neighbourhood, the specificity and generalisation values increase (get worse) with increasing mis-registration.
The results for different sizes of shuffle neighbourhood demonstrate that the range of mis-registration over which distinct values of specificity and generalisation are obtained increases as the neighbourhood size increases.
The results of the comparison between three different methods of NRR are shown in Fig. 15 These show that, particularly in terms of specificity, we can distinguish between the three approaches, with the fully groupwise method performing best, as anticipated. A model built using this approach is shown in Fig. 14.