I found a solution for the use flirt transformation matrices in feat here on the mailing list (https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1005&L=FSL&P=R41795&1=FSL&9=A&I=-3&J=on&X=11FFB324C64E630B8F&Y=truest%40gmx.net&d=No+Match%3BMatch%3BMatches&z=4), hope it may of use to someone.
Unfortunately I still cannot register using a mask using the normmi cost function. I figured it's more appropriate to use the mask on the reference image as it's there where I want to discard the superior and inferior parts.
It's worth saying that using the mask works fine for other cost functions (mutual info, correlation ratio). The normmi cost function without mask yields the best results so far but I'd really like to make it work with the masks, has anyone any suggestions?
Thanks very much in advance for your input.
Cheers,
Torsten
|