PDF version of this entire document

Perturbing Ground Truth

A test set of different registrations was created by applying smooth pseudo-random spatial warps (based on biharmonic Clamped Plate Splines) to each image in the registered set. Each warp was controlled by 25 randomly placed knot-points, each displaced in a random direction by a distance drawn from a Gaussian distribution whose mean controlled the average magnitude of pixel displacement over the whole image. Registration quality was measured, for each level of registration degradation (perturbation), using several variants of each of the proposed assessment methods.

Figure 8: An example brain image and its accompanying anatomical labels, which include the whitematter, graymatter, caudate nucleus, and lateral ventricle

Overall, the above approach was applied 10 times using 10 different random seeds to ensure that both methods are consistent and the results unbiased. The 10 different warp instantiations were generated for each image for each of seven progressively increasing values of average pixel displacement. Figure 9 provided examples from the data as perturbation extent is increased.