PDF version of this entire document

Comparing Registration Algorithms

We compared the results of three registration algorithms as outlined in Section IV-D:

• Pairwise registration of each training set image to a fixed reference image, using an image from the training set as a reference.
• Groupwise registration based on the MDL algorithm described above, with no constraints on the spatial deformations during the registration process (Groupwise 1).
• Groupwise registration based on the MDL algorithm described above, using a statistical shape model to constrain the allowed spatial deformations between the images during registration  [16] (Groupwise 2).

The results are shown in Figure 13. The specificity obtained for the two groupwise methods is significantly better than that obtained using the pairwise approach, implying better registration, but it is not possible to distinguish between the two groupwise methods. As might be expected from the sensitivity results presented above, it is not possible to distinguish between any of the methods using generalisation.

Figure 13: Generalisation and Specificity of the three registration methods as a function of the number of modes included in the appearance model.