open-discussion > problem with bis_combinetransformations
Showing 1-4 of 4 posts
Display:
Results per page:
Apr 15, 2013  03:04 AM | Filipe Monteiro
problem with bis_combinetransformations
Hello,

I have one doubt. I'm trying to perform the ISAS method in a "automatic" script, and now i'm in the part in wich i need to perform
the Inverted transformation (of the non-linear registration).
Using the bis_combinetransformations algorithm, it asks for an input image... what should i put here? shouldn't this method only invert the
previous transformation? (non linear registration). this previous transformation file *.gdr is an input in "--inp2" parameter right? and
i already put the argument "--inverse" to 1... so, do i need anything more? what to insert in the "--inp" argument?

Best Regards,
Filipe monteiro
Apr 15, 2013  05:04 PM | Xenophon Papademetris
problem with bis_combinetransformations
The image is the reference image of the new transformation. So let's assume that you have two transformations

1 Reference MRI->Individual MRI
2 Individual MRI->Individual SPECT

Which when concatenated enable you to reslice the individual spect to the ref MRI.

To invert his combo, you use the individual SPECT as the image as this is now the "new" reference.

Xenios
Apr 15, 2013  06:04 PM | Filipe Monteiro
RE: problem with bis_combinetransformations
Thanks for the quick answer!

So, are you saying that i just need to run:
  • bis_nonlinearintensityregister --inp "interictal image" --inp2 "ref.MRI" --out "outGRD file" ? 
  • Or do I use the bis_combinetransformations --inverse 1 --inp "interictalimage" --inp2 "outGRD" --out "inverse_outGRD" ?

I tried the first way, and the result was a little different than running it through the bioimage suite application itself (the invert button)...
and the 2nd way was a lot more different... it had more than 20MB, and comparing to the file resulting from the bioimage Suite gui method,the header is a lot different.

I understood the idea to invert that combo transformation, but im not getting the results i should (i think)...
Am i doing anything wrong?

Regards,
Filipe Monteiro
Apr 16, 2013  03:04 AM | Filipe Monteiro
RE: problem with bis_combinetransformations
Ok, i think the problem is solved...

I used pxmat_inverttransform.tcl to invert the previous non linear transformation... and obtained the same output as if i inverted from the bioimagesuite GUI.

So i think it is done!

Thanks for your reply though!

Regards,
Filipe Monteiro