Dear all
I am having a similar problem running the tbss_2_reg script. Almost all the
registrations worked correctly, but three did not finish (the source and
target data were different in all three). In each case, there were no
affine.dof and nonlinear.dof files and so an empty .msf file was generated.
In each of these cases, the inverse transformation had worked without any
errors.
I thought it might be a memory problem, but when I reran the script with
just these subjects (after deleting the appropriate .msf files) I got the
following error messages
vertex16 ma601 (../TBSS_GA/tmp): tbss_2_reg
processing 230_TB-NNU38_dti_FAi_to_160_CR1132_dti_FAi
Reading target ... done
Reading source ... done
.
.
.
Iteration = 6 (out of 20), step size = 20
Numerical Recipes run-time error...
Singular matrix in routine ludcmp
...now exiting to system...
Reading target ... done
Reading source ... done
itkRigidTransformation::CheckHeader: Can't open file
230_TB-NNU38_dti_FAi_to_target_affine.dof
itkRigidTransformation::CheckHeader: Can't open file
230_TB-NNU38_dti_FAi_to_target_affine.dof
itkRigidTransformation::CheckHeader: Can't open file
230_TB-NNU38_dti_FAi_to_target_nonlinear.dof
** ERROR: nifti_image_read(grotNx): can't open header file
** ERROR: nifti_image_open(grotNx): bad header info
Error: failed to open file grotNx
Error:: FslGetIntensityScaling: Null pointer passed for FSLIO
** ERROR: nifti_image_read(grotNy): can't open header file
** ERROR: nifti_image_open(grotNy): bad header info
Error: failed to open file grotNy
Error:: FslGetIntensityScaling: Null pointer passed for FSLIO
** ERROR: nifti_image_read(grotNz): can't open header file
** ERROR: nifti_image_open(grotNz): bad header info
Error: failed to open file grotNz
Error:: FslGetIntensityScaling: Null pointer passed for FSLIO
** ERROR: nifti_image_read(grotSx): can't open header file
** ERROR: nifti_image_open(grotSx): bad header info
Error: failed to open file grotSx
Error:: FslGetIntensityScaling: Null pointer passed for FSLIO
** ERROR: nifti_image_read(grotSS): can't open header file
** ERROR: nifti_image_open(grotSS): bad header info
Error: failed to open file grotSS
ERROR: Could not open image grotSS
Image Exception : #22 :: Failed to read volume grotSS
/poppy/homepublic/manjari/fsl/bin/tbss_reg: line 44: 12250 Aborted
$FSLDIR/bin/avwstats++ grotSS -M -P 50 >../${I2R}_nonlinear.msf
Any idea what I'm doing wrong? How easy would it be to register
SubjectA-->SubjectB by generating the inverse of the SubjectB-->SubjectA
transformation dof file? If I do this, how can I find the mean and median
warping in order to create the .msf file?
Thanks for your help,
Mustafa
---------------------------------------------
Mustafa Anjari
Robert Steiner MR Unit
Imaging Sciences Department
Imperial College London
Hammersmith Hospital
Du Cane Road
London W12 0HS
UK
Tel: +44 20 838 32336
Email: [log in to unmask]
On Mon, 1 Jan 2007 22:46:07 +0000, Christian Beckmann
<[log in to unmask]> wrote:
>Hi,
>
>difficult to say - did you maybe run out of disc space? tbss_2_reg is
>just a simple shell script - if you're familiar with shell scripting
>it should be straightforward to sort out how it works. If you want to
>register just a single file the easiest is to simply run the script
>again. If target.hdr exists the script will use the target. Also, it
>will skip all blah_FAi.??? files for which there is a
>blah_to_blah_nonlinear.msf file. Simply remove the msf file for the
>one file that needs re-doing and all should be fine
>cheers
>christian
>
>
>On 1 Jan 2007, at 17:33, Tal Herbsman wrote:
>
>> I just finished up TBSS non linear reg step (tbss_2_reg).
>>
>> For some reason there was not a subject_to_target_FAi_nonlinear.dof
>> file
>> generated for one of the subjects. the .msf file exists and there
>> does not
>> seem to have been a similar issue for any other registration for
>> this or any
>> other subjects (this subject was registered to every other subject
>> just fine).
>>
>> Unfortunately I did not keep logs of the tbss_2_reg step. Any ideas
>> how I
>> can go ahead and debug this problem? Is there a quick way of
>> reregistering
>> just this subject to the specific target.
>>
>> Thanks,
>>
>> Tal.
>>
>> ---------
>> Tal Herbsman, MD
>> Psychiatry
>> Medical University of South Carolina
>
>--
> Christian F. Beckmann
> Oxford University Centre for Functional
> Magnetic Resonance Imaging of the Brain,
> John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
> Email: [log in to unmask] - http://www.fmrib.ox.ac.uk/~beckmann/
> Phone: +44(0)1865 222551 Fax: +44(0)1865 222717
>=========================================================================
|