I think I've found the answer to my question, but hopefully someone on this list can confirm. It seems the 2001 reference for FLIRT[^1] says: "Note that to form the cost functions it is necessary that low values represent good registrations, so many definitions are reversed such as C^MI = -Mutual Information."
This seems to indicate that the actual mutual information value is just the negation of the value FLIRT gives. Correct?
Best,
Andrew Davis
[1] M. Jenkinson and S.M. Smith. A global optimisation method for robust affine registration of brain images. Medical Image Analysis, 5(2):143-156, 2001.
|