Hi
There may be a problem with the affine transform. Could you “cat FA_registered.mat” ?
Note that the code normalises the output vectors internally, so you can’t interpret the norm of the output in terms of the norm of the input.
Cheers
Saad
> On 20 Sep 2017, at 17:07, Max Wichmann <[log in to unmask]> wrote:
>
> Dear Fsl Team,
> when I try to register my direction results (calculated with another program (Mrtrix)) to a template I got following error message:
>
> vecreg -i peak1_reorientated.nii.gz -o peak1_reorientated_registered.nii.gz -r FA_mean.nii.gz -t FA_registered.mat
> Registering vector...
> libc++abi.dylib: terminating with uncaught exception of type NEWMAT::ConvergenceException
> Abort trap: 6
>
> I guess it has something to do with the fact that the orientations are not calculated via fsl. But the direction files have the same structure as the dyad files! 4d files with 1dim=x 2dim=y 3dim=z and 4dim=1st. volume x-coordinate of given vector; 2nd. volume y-coordinate of given vector; 3rd volume z coordinate of given vector! They are able to load into fsleyes without any problems! The only differences compared to the dyads is that the orientations are not normalised to 1.
>
> It would be so nice if somebody understand the error message and could help me to fix the problem!
> Thank you very much.
>
> Best
> Max
|