# About calculating FA(Fractional Anisotropy) with TensorRelativeAnisotropyImageFilter

**URL:** https://discourse.itk.org/t/about-calculating-fa-fractional-anisotropy-with-tensorrelativeanisotropyimagefilter/7437
**Category:** Beginner Questions
**Tags:** itk
**Created:** [February 18, 2025, 8:07am UTC](https://discourse.itk.org/t/about-calculating-fa-fractional-anisotropy-with-tensorrelativeanisotropyimagefilter/7437 "2025-02-18T08:07:08Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![Pinyu\_Huang](https://discourse.itk.org/letter_avatar_proxy/v4/letter/p/34f0e0/32.png) [@Pinyu\_Huang](https://discourse.itk.org/u/Pinyu_Huang)
#### Post date: [February 18, 2025, 8:07am UTC](https://discourse.itk.org/t/about-calculating-fa-fractional-anisotropy-with-tensorrelativeanisotropyimagefilter/7437/1 "2025-02-18T08:07:08Z")

</div>

Hi there, I tried to calculate FA and PDD for streamlines tractography with TensorRelativeAnisotropyImageFilter.

And I basically follow the workflow in the website ([ITK: Examples/Filtering/DiffusionTensor3DReconstructionImageFilter.cxx](https://docs.itk.org/projects/doxygen/en/stable/Examples_2Filtering_2DiffusionTensor3DReconstructionImageFilter_8cxx-example.html)), But as a result, the FA image and colored-FA image seems a little bit noisy and mosaic.

 ![image](https://discourse.itk.org/uploads/default/original/2X/1/1f9845329d587dc810998ccc4f63f9fdd4fca028.png) ![image](https://discourse.itk.org/uploads/default/original/2X/c/cd0e8729e647dae8401fbc4085684b6fe6106dd3.png)

While it should look like this:

 ![image](https://discourse.itk.org/uploads/default/original/2X/3/30f475d2d118477fc5d9da40cedc25fbce91cc22.jpeg)

could you please give me some advice?

I think there’re maybe some reasons:

1. some preprocessments and postprocessments I skipped to make FA more good-looking
2. some settings in `TensorRelativeAnisotropyImageFilter` should be set but I forgot
3. before calculating FA and PDD I should up-sample the image, but… how big is reasonable?
4. …etc? I have no idea…

Great thanks for any help!