Hello, we are studying a quantitative measure of MR images distortion (in
particular, spinal cord EPI). We are thinking of using the average Jacobain
determinant to account for "zooming", and the deformation field of a warping
between distorted and undistrorted image for accounting for geometrical
distortions mantainig the volume.
However, and unexpectedly, the average value of the components of a
deformation field for images registered and with very few distortion is positive,
while the average jacobian is, as expected, near 1. Please see in attachment a
2-D vector plot of the deformation field of a spoecific slice (colors map the
absolute value of the 3d vector). Any hints?
Bestm, Federico
|