Hi,
Your argument to -a needs to be a flirt matrix, not an image.
This is why everything to do with the transformation of the
priors is failing. Just register your image to the standard
space and then pass that flirt matrix to fast with the -a
option and it should work fine.
All the best,
Mark
On 7 Aug 2009, at 03:16, Yannis Paloyelis wrote:
> Dear all,
>
> I try to use FAST using priors to aid segmentation. I am getting some
> worrying messages while the process runs though - is it likely to be a
> problem? (e.g. see last three lines)
>
> Thanks so much for your help!
>
> (I use an intel mac with os x 10.5 and the latest fsl package).
>
> the command I use:
>
> fast -g -a ./flirt_pre/10_brain8_brain_std_pre.nii.gz -b -B -o
> ./fast_prior/fastpr -v -p 10_brain8_brain &
>
> and these are the comments I get:
>
> /Applications/fsl/bin/flirt -ref 10_brain8_brain -in
> /Applications/fsl/data/standard/tissuepriors/avg152T1_csf -out
> ./fast_prior/fastpr_csf_stdspace -applyxfm -init
> ./flirt_pre/10_brain8_brain_std_pre.nii.gz
> terminate called after throwing an instance of
> 'NEWMAT::NotSquareException'
> /Applications/fsl/bin/flirt -ref 10_brain8_brain -in
> /Applications/fsl/data/standard/tissuepriors/avg152T1_gray -out
> ./fast_prior/fastpr_gm_stdspace -applyxfm -init
> ./flirt_pre/10_brain8_brain_std_pre.nii.gz
> terminate called after throwing an instance of
> 'NEWMAT::NotSquareException'
> /Applications/fsl/bin/flirt -ref 10_brain8_brain -in
> /Applications/fsl/data/standard/tissuepriors/avg152T1_white -out
> ./fast_prior/fastpr_wm_stdspace -applyxfm -init
> ./flirt_pre/10_brain8_brain_std_pre.nii.gz
> terminate called after throwing an instance of
> 'NEWMAT::NotSquareException'
>
> csf prior image not transformed correctly! priors are not used!
> Beginning prior-based initialisation
>
|