[Insight-developers] NRRD Image Format (and tensor estimation thereafter?)

[Karthik Krishnan]

Gordon Kindlmann wrote:

> hello,
>
> Which version of ITK are you using?
> The most general handling of NRRD files was added by me prior to the  
> 2.2 release.
>
> Are you reading the file as a scalar image?
> I am guessing this may be the problem.
>
> I believe you can follow the example of:
> Insight/Testing/Code/IO/itkNrrdVectorImageReadTest.cxx
>
> which uses a itk::Vector<float, N> PixelType for which the length N  
> is known at compile time,
>
> or the example of:
> Insight/Testing/Code/Common/itkVectorImageTest.cxx
>
> which uses VectorImage<float, D> image type for which the image  
> dimension D is known at compile time, but the length (number of  
> values per sample) can be set at run-time, based on a read file.
>
> However, others more knowledgeable than I can chime in with further  
> details on the use of VectorImage, also new with ITK 2.2.
>
> Once you have DWIs loaded in, you probably want to do tensor  
> estimation.  I am not sure what the current status of ITK-based  
> tensor estimation is, specifically, can it work with VectorImages?
>
The tensor framework in ITK usually has the following pipeline:

Reference + DWI Gradient images  ----> 
DiffusionTensor3DReconstructionImageFilter  -----> itk::Image< 
DiffusionTensor3D< TScalar >, 3 >

To this image you can apply one of the may tensor filters like:

TensorFractionalAnisotropyImageFilter,
TensorRelativeAnisotropyImageFilter,
Hessian3DToVesselnessMeasureImageFilter,
SymmetricEigenAnalysisImageFilter,

I presume your question was "Can the DWI Gradient images" be a 
itk::VectorImage.
Not with the way the filter is written right now, It expects 'n' 
itk::Images as input, but its very easy to make it expect an 
itk::VectorImage as input. If you want, I can change that or add another 
filter.

Would you rather have it expect a VectorImage ?

Regards
karthik

> Gordon
>
>
> On Dec 1, 2005, at 10:50 AM, Vincent A. Magnotta wrote:
>
>> I am trying to use the NRRD format to store diffusion weighted  
>> images. I
>> have been using the information provided on the NA-MIC website:
>>
>> http://www.na-mic.org/Wiki/index.php/NAMIC_Wiki%3ADTI%3ATensor_format
>>
>> It appears that the ITK implementation for diffusion weighted images
>> does not work properly. From just a first pass analysis, it seems that
>> if the NRRD file describes an image that is 4D, but it only has three
>> spatial dimensions, the 4th dimension upon loading the images has a  
>> size
>> of only 1. I have a very small image example that illustrates the
>> problem that is attached. When I print out the image after loading  
>> it, I
>> get the following information:
>>
>>
>> Image (0x84e5bd8)
>>   RTTI typeinfo:   itk::Image<short, 4>
>>   Reference Count: 2
>>   Modified Time: 131
>>   Debug: Off
>>   Observers:
>>     none
>>   Source: (none)
>>   Source output index:  0
>>   Release Data: Off
>>   Data Released: False
>>   Global Release Data: Off
>>   PipelineMTime: 11
>>   UpdateMTime: 130
>>   LargestPossibleRegion:
>>     Dimension: 4
>>     Index: [0, 0, 0, 0]
>>     Size: [2, 2, 2, 1]
>>   BufferedRegion:
>>     Dimension: 4
>>     Index: [0, 0, 0, 0]
>>     Size: [2, 2, 2, 1]
>>   RequestedRegion:
>>     Dimension: 4
>>     Index: [0, 0, 0, 0]
>>     Size: [2, 2, 2, 1]
>>   Spacing: [2, 2.5, 2, 1]
>>   Origin: [0, 0, 0, 0]
>>   Direction:
>> -1 0 0 0
>> 0 -1 0 0
>> 0 0 1 0
>> 0 0 0 1
>>
>> I will look into this some more over the next couple of days, but I
>> thought that I would post it here as well. I am just getting familiar
>> with the format, so if anyone else has suggestions they would be
>> helpful.
>>
>> Vince
>>
>>
>>
>> -- 
>> Assistant Professor
>> Department of Radiology
>> 0453-D JCP
>> 200 Hawkins Drive
>> Iowa City, IA 52242
>> E-mail: vincent-magnotta at uiowa.edu
>> Phone: 319-356-8255
>> Fax: 319-353-6275
>> Website: http://www.radiology.uiowa.edu
>> <tmp.img>
>> <tmp.nhdr>
>> _______________________________________________
>> Insight-developers mailing list
>> Insight-developers at itk.org
>> http://www.itk.org/mailman/listinfo/insight-developers
>
>
> _______________________________________________
> Insight-developers mailing list
> Insight-developers at itk.org
> http://www.itk.org/mailman/listinfo/insight-developers
>