Comparison of patient dataset 201 and 301
In order to compare the two dataset we concentrated on the T2 weighted
images that correlate with anatomy. We performed first a rigid registration
in order to correct for the patient position. In a second run we corrected
for non-rigid transformations that should appear if there are sufficiently
large anatomical differences between the data sets.
The method used analyses the displacement vector field which is the result
of the non-rigid registration. The rate of strain tensor is correlated with
the stress in the material and therefore provides information about the
possible change between the two scans. Note that the rate of strain tensor
is computed on the displacement vector fields of the registration only thus
does not provide information about the water diffusion in the tissue.
Results
Single images
The rate of strain tensor was analysed to obtain a regional distribution of expansion (depicted red) and compression (depicted blue) in the data. The colormap either codes for the magnitude of the vector field or, in case of the EllipseDisplay for the trace of the tensor.
A series of color overlays. We used the b=0 - T2 weighted images and fused
them with the scalarfield containing the trace of the tensor.