Adding new filtering script by orientation

[EmmaRenauld]

New script to filter a tractogram based on total changes in x, y or z directions. There is a note explaining that x, y, z mean the x,y,z coordinates of the sft streamlines, not necessarily aligned with the real x,y,z coordinates of the brain.

This will be useful to separate cingulum streamlines from the CC in our new tractometer based on ROI, using scil_score_tractogram (upcoming).

[pep8speaks]

Hello @EmmaRenauld, Thank you for updating !

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2022-05-05 20:18:25 UTC

[EmmaRenauld]

@mdesco @AntoineTheb

[scil-admin]

Build passed ! Good Job 🍻 !

[scil-admin]

Build passed ! Good Job 🍻 !

[scil-admin]

Build passed ! Good Job 🍻 !

[AntoineTheb]

Just finished looking at the code and running tests, LGTM ! Just one minor comment in the code. Do you have data to test it ? (with expected results). Thanks !

[EmmaRenauld]

You can check it with the CC (https://drive.google.com/file/d/1JxC8PfK9DfzPYiuODuzAtunVj1JpQwH-/view?usp=sharing):

scil_filter_streamlines_by_orientation.py CC_u_shaped.trk tmp.trk -v
scil_filter_streamlines_by_orientation.py CC_u_shaped.trk tmp2.trk -v --use_abs

The first version will show you that the minimum x distance is 0, because some streamlines have a very pronounced U shape, to the point of nearly coming back to their initial position. The second one will show you that the minimum x distance is 43 because those streamlines did actually do a loop, they did not simply go from point 0 to point 0 by not moving.

(In the ISMRM, first version will also show you that y distance is max 31.9mm, showing a good separation with, ex, the cingulum, which would have max distance of ~127).

[scil-admin]

Build passed ! Good Job 🍻 !

[scil-admin]

Build passed ! Good Job 🍻 !

[mdesco]

Minor edits and then all good

[scil-admin]

Build passed ! Good Job 🍻 !