[MINC-users] mincresample apply nonlinear transformation using .xfm with Transform Type = Grid_Transform

[Alex Zijdenbos]

'minccomplete' has two 'c's :)

you get the same error for all grid_0.mnc files, or just one?

my other guess would be you are looking at a MINC1/MINC2 issue, and/or
that you might have an ancient MINC install. What does 'mincresample
-version' tell you?

On Thu, Jun 20, 2013 at 3:31 PM, Sébastien Proulx
<sebastien.proulx2 at mcgill.ca> wrote:
> It seems like I don't have access to minccomplete, but see the following
> terminal output:
> -------------------------------------------------------------------------------------
> duncan:~/transformTest_bis/Transforms_45/nonlinear> mincomplete
> TMS_CON_45_nlfit_It_grid_0.mnc
> mincomplete: Command not found.
> duncan:~/transformTest_bis/Transforms_45/nonlinear> mincinfo
> TMS_CON_45_nlfit_It_grid_0.mnc
> ncopen: filename "TMS_CON_45_nlfit_It_grid_0.mnc": NetCDF: Unknown file
> format
> miopen: MINC package entry point
> -------------------------------------------------------------------------------------
>
> Since I tried with two different subjects, I thought a corruption was
> unlikely, but I guess mincinfo should be able to read any intact minc file.
> Is it possible that minctracc (see content of .xfm) did correctly produce
> the nonlinearly registered images, but somehow screwed-up all the
> ..._nlfit_It_grid_0.mnc files?
>
> Thanks a lot Alex!
>
>
> Sébastien Proulx, MSc
> Research Assistant
> McGill University
> (514) 398-6644 #00445
>
> Le 13-06-20 2:22 PM, Alex Zijdenbos a écrit :
>>
>> Hi Sébastien,
>>
>> My guess would be that the deformation field (grid_0.mnc) file is
>> corrupt. What happens when you run this:
>>
>> minccomplete TMS_CON_44_nlfit_It_grid_0.mnc
>>
>> and/or mincfinfo?
>>
>> -- A
>>
>> On Thu, Jun 20, 2013 at 2:00 PM, Sébastien Proulx
>> <sebastien.proulx2 at mcgill.ca> wrote:
>>>
>>> Hello Minc people,
>>>
>>> My colleague gave me transformation files from a nonlinear registration
>>> to a
>>> template done with minctools. I'm trying to use those to bring some
>>> images
>>> (segmented in MNI space) back to their native space. Linear transforms
>>> are
>>> working fine, but I apply the inverse of the nonlinear part:
>>>
>>> mincresample -inverse_transformation -transformation
>>> nonLinTransformFileName.xfm -like nativeSpaceAnatomic.mnc
>>> templateSpaceSegmentedImage.mnc nativeSpaceSegmentedImage.mnc
>>>
>>>
>>> doesn't work whatever I do (reloading files or try simple variants of the
>>> command):
>>>
>>> Error: opening MINC file "./TMS_CON_44_nlfit_It_grid_0.mnc".
>>> input_transform: error reading transform.
>>> Error reading transformation file.
>>>
>>>
>>> Here is the content of nonLinTransformFileName.xfm:
>>>
>>> ---------------------------------------------------------------------------------------------------
>>> MNI Transform File
>>> %Sun Aug  5 12:30:27 2012>>> minctracc -clobber -nonlinear corrcoeff
>>> -weight
>>> 1 -stiffness 1 -similarity 0.3 -sub_lattice 6 -iterations 10 -step 4 4 4
>>> -lattice_diam 12 12 12 -transformation
>>> /tmp/best1stepnlreg.pl-UAIGq86o/STMS_CON_44_t1_final_4.xfm -source_mask
>>>
>>> /sb/project/eim-670-ab/gridshare/15/52/01/stremblay-Civet-T155201/civet_out/CON_44/mask/TMS_CON_44_skull_mask.mnc
>>> -model_mask
>>>
>>> /sb/home/prioux/tools/CIVET/Feb-2010-a/share/mni-models//icbm_avg_152_t1_tal_nlin_symmetric_VI_mask.mnc
>>> /tmp/best1stepnlreg.pl-UAIGq86o/STMS_CON_44_t1_final_2_blur.mnc
>>>
>>> /tmp/best1stepnlreg.pl-UAIGq86o/Ticbm_avg_152_t1_tal_nlin_symmetric_VI_2_blur.mnc/sb/project/eim-670-ab/gridshare/15/52/01/stremblay-Civet-T155201/civet_out/CON_44/transforms/nonlinear/TMS_CON_44_nlfit_It.xfm
>>> %(Package mni_autoreg 0.99.3, compiled by tnguyen at localhost.localdomain
>>> (x86_64-unknown-linux-gnu) on Mon Aug 30 17:14:08 EDT 2010)
>>>
>>> Transform_Type = Linear;
>>> Linear_Transform =
>>>   1 0 0 0
>>>   0 1 0 0
>>>   0 0 1 0;
>>> Transform_Type = Grid_Transform;
>>> Displacement_Volume = TMS_CON_44_nlfit_It_grid_0.mnc;
>>>
>>> ---------------------------------------------------------------------------------------------------
>>>
>>>
>>> Is there something I'm missing, like non linear transformations use
>>> another
>>> command? Note that I cannot open TMS_CON_44_nlfit_It_grid_0.mnc with
>>> register either, but I bet this is expected...
>>>
>>> Thanks a lot for your help, I really have no idea what to do with that...
>>> apart from rerunning the segmentation in native space!
>>>
>>> Sébastien Proulx, MSc
>>> Research Assistant
>>> McGill University
>>> (514) 398-6644 #00445
>>>
>>> _______________________________________________
>>> MINC-users at bic.mni.mcgill.ca
>>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>>>
>> _______________________________________________
>> MINC-users at bic.mni.mcgill.ca
>> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>>
>>
>
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>