Print

Print


Hi, Mark

As you say, my error was due to an initial problem with the 4.1.5 after release.
I downloaded it on 26 Jan 2010.

When I redownloaded fsl-4.1.5 yesterday and installed it on my mac pro, 
I could execute tbss_deproject without trouble.

In this mailing list, there were some error reports about invwarp after release.
The subject name is "Bug in invwarp Wrapper" and "Inwarp error in new FSL release".
My error also might be the same reason. 

For your information:
When I was running tbss_deproject (I downloaded on 26 Jan 2010),
the error message is below; 

Thanks,
Kenji

********************************************************
tbss_deproject grot 2
de-projecting <skeleton-space-image> into all_FA (nonlinearly aligned) subject space
output from stage 1 deprojection will be grot_to_all_FA
de-projection done
now creating inverse warp fields...

list of original images = Con_16473_FA Con_25260_FA......
original target = target
processing Con_16473_FA:
***************************************************
The following COMPULSORY options have not been set:
	-r,--ref	filename for reference volume
***************************************************

Part of FSL (build 415)
fnirtfileutils
Copyright(c) 2007, University of Oxford (Jesper Andersson)

Usage: 
fnirtfileutils --in=my_fnirtcoefs --ref=my_refvol --out=my_field
fnirtfileutils --in=my_fnirtcoefs --ref=my_refvol --jac=my_jacobians
fnirtfileutils --in=my_fnirtcoefs --ref=my_refvol --out=my_field --withaff


Compulsory arguments (You MUST set one or more of):
	-i,--in		filename of input coefficient volume (to be converted)
	-r,--ref	filename for reference volume

Optional arguments (You may optionally specify one or more of):
	-o,--out	filename for output (field/coef) volume
	-f,--outformat	Output format [field spline], default=field
	-w,--warpres	Warp resolution (mm), only relevant when --outformat=spline
	-k,--knotspace	Knot-spacing (voxels), only relevant when --outformat=spline
	-j,--jac	filename for output (jacobian map) volume
	-a,--withaff	If set, the affine transform is included in the field/jacobian
	-v,--verbose	switch on diagnostic messages
	-h,--help	display this message


** ERROR (nifti_image_read): failed to find header file for 
'/tmp/fsl_mNRTFfA/hr2std_field'
** ERROR: nifti_image_open(/tmp/fsl_mNRTFfA/hr2std_field): bad header info
Error: failed to open file /tmp/fsl_mNRTFfA/hr2std_field
Cannot open volume /tmp/fsl_mNRTFfA/hr2std_field for reading!
...
...
*******************************************************************

On Tue, 9 Feb 2010 15:45:06 +0000, Mark Jenkinson <[log in to unmask]> wrote:

>Hi,
>
>This actually isn't that helpful as we still don't know
>what goes wrong for you before you make any changes.
>We do not have problems here, and you should definitely
>not need to change invwarp for invwarp_exe.
>
>There was an initial problem with the 4.1.5 patch, so if
>you picked it up in the first few weeks after release then
>maybe check to see if the latest version still causes you problems.
>
>But either way it would help us greatly if you could describe
>in more detail what goes wrong.  What are the errors?
>What files are produced?  What conditions does it occur in?
>
>All the best,
>	Mark
>
>
>On 9 Feb 2010, at 10:22, Kenji wrote:
>
>> Hi, Mark.
>>
>> Though FSL4.1.4 has one invwarp in /fsl/bin file ,i.e "invwarp",
>> FSL4.1.5 has two invwarp , i.e "invwarp" and "invwarp_exe".
>>
>> When I exchanged "invwarp" for "invwarp_exe" in
>> tbss_deproject(FSL4.1.5) as
>> below,
>> tbss_deproject was executed without trouble.
>>
>> You might understand this difference.
>>
>> ** tbss_deproject(FSL4.1.5) **
>>     ......
>>    for g in $list ; do
>> 	echo "processing ${g}:"
>> 	nextvol=`$FSLDIR/bin/imglob ../stats/${IN}_to_all_FA_tmp0* | awk
>> '{print $1}'`
>> 	warp=${g}_to_${target}_warp
>> 	if [ `$FSLDIR/bin/imtest ${warp}_inv` = 0 ] ; then
>> 	   # $FSLDIR/bin/invwarp -w $warp -o ${warp}_inv -r $g
>>              $FSLDIR/bin/invwarp_exe -w $warp -o ${warp}_inv -r $g
>> 	fi
>>    ......
>>
>> I hope this is useful.
>> kenji
>>
>>
>> On Mon, 8 Feb 2010 21:33:43 +0000, Mark Jenkinson
>> <[log in to unmask]> wrote:
>>
>>> Hi,
>>>
>>> We are unaware of any unresolved bugs with invwarp and tbss.
>>> Can you please be more specific about what happens?
>>>
>>> All the best,
>>> 	Mark
>>>
>>>
>>> On 8 Feb 2010, at 19:04, Kenji wrote:
>>>
>>>> Hi FSL experts,
>>>>
>>>> I have recently discovered a bug in the "tbss_deproject grot 2" in
>>>> FSL 4.1.5.
>>>> I think this error might be due to a bug in the invwarp script.
>>>>
>>>> When I use the invwarp in FSL4.1.4 instead of it in FSL4.1.5,
>>>> I could execute "tbss_deproject grot 2" in FSL4.1.5.
>>>>
>>>> Maybe, has the bug already been corrected ?
>>>>
>>>> Kenji
>>>>
>>