Applying image-based deformations to streamlines example

[BramshQamar]

Description

In this example on applying image-based deformations to streamlines doc/examples/streamline_registration.py,
I don't understand what this part is doing

dipy/doc/examples/streamline_registration.py

Lines 32 to 34 in 541b5b2

	if not op.exists('lr-superiorfrontal.trk'):
	from streamline_tools import *
	vox_size = hardi_img.header.get_zooms()[0]

Also gives me an error when I run it:

ModuleNotFoundError: No module named 'streamline_tools'

And why are we accessing hardi_img variable before it is initialized?

dipy/doc/examples/streamline_registration.py

Line 34 in 541b5b2

vox_size = hardi_img.header.get_zooms()[0]

[arokem]

Hi Bramsh, Fair question. It's a bit of a strange pattern: the example you are trying to run depends on running this other example first: https://github.com/dipy/dipy/blob/541b5b2992e786b5f51d05649a398cb331903456/doc/examples/streamline_tools.py That's the thing it's trying to import. That also explains where the already-initialized variable is coming from. But it also means that the example will only work if it is run in the context of our source tree. This is not a free-standing example that runs on its own. Maybe we should document this better?

…

On Thu, May 20, 2021 at 9:14 AM Bramsh Q Chandio ***@***.***> wrote: Description In this example on applying image-based deformations to streamlines doc/examples/streamline_registration.py <http://url>, I don't understand what this part is doing https://github.com/dipy/dipy/blob/541b5b2992e786b5f51d05649a398cb331903456/doc/examples/streamline_registration.py#L32-L34 Also gives me an error when I run it: ModuleNotFoundError: No module named 'streamline_tools' And why are we accessing hardi_img variable before it is initialized? https://github.com/dipy/dipy/blob/541b5b2992e786b5f51d05649a398cb331903456/doc/examples/streamline_registration.py#L34 — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2400>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAA46NUI2QMOJ5C37OHI6PTTOUYNLANCNFSM45HLIFOA> .

[BramshQamar]

Hi @arokem

Thank you for the explanation. Running this tutorial https://github.com/dipy/dipy/blob/541b5b2992e786b5f51d05649a398cb331903456/doc/examples/streamline_tools.py solved the issues.
It would be helpful to add prerequisites to the documentation.

However, I found out that the output generated by https://dipy.org/documentation/1.4.1./examples_built/streamline_registration/#example-streamline-registration is very strange looking.

Here green is the original bundle and pink after registration.

[arokem]

Might be worth viewing in the context of the MNI template. It might have moved to the right place with respect to the template?

[BramshQamar]

Visualized with template

[arokem]

Yeah - that's no good.

[BramshQamar]

I think the issue is that moving bundle and image are not in the same space to begin with

So when static and moving images are aligned and that transformation is applied to the bundle it generates spurious output as that bundle was not in the same space as the moving image. It is assumed that moving image and bundle are in the same space, probably coming from the same subject which doesn't seem to be true in this case.

[BramshQamar]

Here's the issue, density map, and the streamlines saved from this example https://github.com/dipy/dipy/blob/541b5b2992e786b5f51d05649a398cb331903456/doc/examples/streamline_tools.py are not in the same space.

[skoudoro]

Thank you for pointing that out. it would be good to:

• separate these 2 tutorials
• fix the space problem

[arokem]

Maybe we should make a fetcher for these trk files, just so we can make this example freestanding?