[Insight-users] DTI data.. can't read

[Torsten Rohlfing]

Gordon:

Thank you for the new header; the geometry is now read okay. However, I 
seem to be getting all "empty" tensors. I will need to dig deeper into 
the data structures to see whether this is an issue with my processing 
(I am generating and writing an FA map), or whether it's an import problem.

I may need to get back to you on this one.

Best,
  Torsten

Gordon Kindlmann wrote:

> hello,
>
>> Hi Gordon --
>>
>> I have a question regarding that DTI data you are making available  
>> through the server at Utah.
>
>
> (for the interested reader this is here: http://www.sci.utah.edu/~gk/ 
> DTI-data/)
>
>> It seems that the NRRD reader currently in ITK cannot actually read  
>> this file. More precisely, it seems that the spatial dimensions are  
>> not correctly separated from the matrix element dimension. Here's a  
>> printout of the resulting image object that may clarify the issue:
>
>
> No, ITK can't usefully read that file, other than to see it as a 4-D  
> volume.  The NRRD0001 version of the format didn't have a way of  
> tagging axes with the kind of information that was stored along it.   
> NRRD0004 and NRRD0005 do have this information, and the ITK nrrd  
> reader goes to some lengths to correctly map the information to  
> current ITK pixel types.
>
>
>> Interestingly, when I create a tensor image within ITK and write it  
>> to a NRRD0004 file, I can read everything fine, so maybe all that's  
>> needed to fix the issue practically would be if you could come up  
>> with a header for the file that can be read correctly by what's in  ITK?
>
>
> Done, sorry for the delay:
> http://www.sci.utah.edu/~gk/DTI-data/gk2/gk2-rcc-mask2.nhdr
>
> Can you make sure this works as expected?
>
> Gordon



-- 
Torsten Rohlfing, PhD          SRI International, Neuroscience Program
 Research Scientist             333 Ravenswood Ave, Menlo Park, CA 94025
  Phone: ++1 (650) 859-3379      Fax: ++1 (650) 859-2743
   torsten at synapse.sri.com        http://www.stanford.edu/~rohlfing/

     "Though this be madness, yet there is a method in't"