Hello Everyone,
I have a quick question. I am using FSL 5.0.4.
I am using "mcflirt" with "-meanvol (register time-series to mean volume)" option to register my time-series images in a contrast enhancement study and it seems that using mean volume gives better registered images. Also, I am using "-mats" option to save transformation matrices (I need them because I have prior drown VOI's).
When I checked the first one's transformation matrix (MAT0000), it is an 4x4 identity matrix (same with different data), which is unusual. Because when I compared the first time's registered and unregistered images, they are different.
Also I am aware that "-meanvol" with "-plot" option had some issues
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0912&L=FSL&P=R34624&1=FSL&9=A&I=-3&J=on&d=No+Match%3BMatch%3BMatches&z=4
Is it a bug or I am missing something? Like all transformation matrixes are referenced to first transformed image? but then how can I find initial transformation matrix in the first place?
Thanks in advance
Volkan
|