[Camino-users] tensor format and get tensor components

[Philip A Cook]

If you call dt2nii, you will get tensors that can be read by DTI-TK. You can do further pre-processing (like scale the tensors) on the DTI-TK side.

To answer your question, if you call dt2nii and then split the resulting dt.nii.gz file, the order would be 

0001	dxx
0002	dxy
0003	dyy
0004	dxz
0005	dyz
0006	dzz

that's the definition of lower-triangular ordering.


> On Aug 5, 2016, at 7:24 PM, xiaoping qu <quxiaoping1 at gmail.com> wrote:
> 
> Hi Philip,
> 
> Thanks for your reply  and it is really helpful.
> I have use split4dnii to split NIFTI format tensor to 6 components (lower triangle), and they are automatically named as
> 0001.nii.gz  0002.nii.gz  0003.nii.gz  0004.nii.gz  0005.nii.gz  0006.nii.gz.
> 
> To use DTITK function to fit tensor, I need to know the what the number represent.
> 
> Does 0001 represent dxx, 
>      0002 represent dxx, 
>      0003 represent day, 
>      0004 represent dzx,
>      0005 represent dzy,
>      0006 represent dzz ?
> 
> The meaning for the tensor components is very important for me to do next process.
> 
> Thanks for your patience and help,
> Xiaoping
> 
> 
> 
> 2016-08-05 15:17 GMT-07:00 Philip A Cook <cookpa at mail.med.upenn.edu>:
> Hi,
> 
> 1. Camino's raw data format is upper triangular, but dt2nii will write tensors with lower triangular elements, as specified by NIFTI. These should work with DTI-TK. You will have to do other pre-processing, see the DTI-TK documentation here
> 
> http://dti-tk.sourceforge.net/pmwiki/pmwiki.php?n=Documentation.BeforeReg
> 
> 
> 
> 2. If you want the components back you can call split4dnii, which will split 4D or 5D nii into 3D images.
> 
> 
> 3. I don't recommend using analyzedti or any Analyze files, they are obsolete now. With NIFTI:
> 
> dtfit data.nii.gz mask.nii.gz | dt2nii -header mask.nii.gz -outputroot myDT_
> 
> This will make
> 
> myDT_dt.nii.gz
> myDT_exitcode.nii.gz
> myDT_lnS0.nii.gz
> 
> You can then do
> 
> fa -inputfile myDT_dt.nii.gz -outputfile myDT_fa.nii.gz
> md -inputfile myDT_dt.nii.gz -outputfile myDT_md.nii.gz
> 
> If you want the full eigen system, things are a little trickier, but you can use voxel2image to get the output of dteig as a series of NIFTI images (see http://camino.cs.ucl.ac.uk/index.php?n=Main.Fileformats). It would be nice to have a dteig2nii to automate this, but we don't currently have that.
> 
> If you are using DTI-TK, then you can take the NIFTI DT and use it's TVTool to get the eigen system, as well as AD, RD, MD, FA, etc.
> 
> 
> 
> 
> > On Aug 5, 2016, at 5:33 PM, xiaoping qu <quxiaoping1 at gmail.com> wrote:
> >
> > Hi Camino experts,
> >
> > I am Xiaoping, a new user to Camino.
> >
> > Since I want the diffusion tensor  components to do registration with DTITK, I am very interested in tensor type and how to abstract tensor components from a tensor file.
> >
> > 1. I tried to use dtfit to fit a tensor with Bdouble format and can use dt2nii to convert Bdouble format file to nifti format.
> > According your on line documents, I find that the tensor.Bdouble from dtfit is upper triangle but no idea about the nifti tensor format?
> > Can you tell me  that after dt2nii, what type the nifty format tensor is ? upper triangle or lower triangular?
> >
> >
> > 2. If I want to extract tensor components like dxx, dxy, dxz, dyy, dzz from tensor.nii.gz, what kind of function I can use?
> >
> >
> >
> > 3. Aso, according to the online documents about camino, I just find analyzedti is a command to get individual analyze format components. If I have the DWI nifti.gz format image and its binary mask(nii,gz) and scheme file. How can I use analyzedti to get individual tensor components?
> >
> > Do I need to fist img2voxel to convert dwi to voxel order as Bfloat file,
> > and then use voxel2scanner to convert dwi.Bfloat to scanner order as Bdouble file, and then use AnalyzeHeader to convert Bdouble to analyze format, and finally use Analyzedti to get analyze format tensor components?
> >
> >
> >
> > Thanks,
> > Xiaoping
> > _______________________________________________
> > Camino-users mailing list
> > Camino-users at www.nitrc.org
> > http://www.nitrc.org/mailman/listinfo/camino-users
> 
>