ENH: Support regularization transform hardening in DICOM Scalar plugin

[Punzo]

this PR adds:

• Grouping frames by orientation with a 10^-6 tolerance in examineFiles in DICOM module.
• Add harden transform option in the DICOM settings for load in DICOMScalarVolumePlugin.

[pieper]

Thanks for working on this 👍

Another stylistic comment is that the commit message is longer than recommended.

https://slicer.readthedocs.io/en/latest/developer_guide/contributing.html#how-to-write-commit-messages

[Punzo]

Thanks for working on this 👍

Another stylistic comment is that the commit message is longer than recommended.

https://slicer.readthedocs.io/en/latest/developer_guide/contributing.html#how-to-write-commit-messages

Thanks for the review Steve. I have addressed all the comments.

@pieper @lassoan up to you to merge if you think it is ready.

[pieper]

Sorry for all the nitpicks : )

But the diffusion issue is important since I don't think the current code will work.

[pieper]

Ahhh, looks nice now : )

[Punzo]

Ahhh, looks nice now : )

perfect, thanks for the review :)

[Punzo]

@jcfr @lassoan @pieper thank you very much for the reviews and sorry for the noise/spam.

I have reverted the class implementation commits. Please have a look at the current commit 0d2ae26 and let me know if there is any another issue.

[jcfr]

Suggested commit message:

ENH: Add DICOM grouping frames by orientation and add harden transform option

Update the ScalarVolume DICOM plugin adding:
* subseries grouping by orientation 
* harden transform option in the DICOM settings.

[Punzo]

Suggested commit message:

ENH: Add DICOM grouping frames by orientation and add harden transform option

Update the ScalarVolume DICOM plugin adding:
* subseries grouping by orientation 
* harden transform option in the DICOM settings.

@jcfr ok, tomorrow I will squash and change the commit message, thanks!

regarding the eps vs self.epsilon, I don't have strong preferences. However the default value of self.epsilon is 0.01. While we need a default value of 1.e-6 for the orientation check. We could multiply self.epsilon for a factor in examineFiles. Up to you.

[jcfr]

To improve on #7362 (review):

ENH: Support regularization transform hardening in DICOM Scalar plugin

Update the logic identifying sub-series (each loaded as independent
volume node) to support grouping by orientation and add harden a new
DICOM settings option to enable automatic hardening of the corresponding
regularization transform.

I think this better describe what was implemented and help understand why.

re: Documentation

Consider also updating:

Slicer/Docs/user_guide/modules/dicom.md

Lines 184 to 187 in ba80bdd

	- DICOMScalarVolumePlugin settings:
	- You can choose what back-end library to use (currently GDCM, DCMTK, or GDCM with DCMTK fallback with the last option being the default. This is provided in case some data is unsupported by one library or the other.
	- Acquisition geometry regularization option supports the creation of a nonlinear transform that corrects for things like missing slices or gantry tilt in the acquisition. See more information [here](https://github.com/Slicer/Slicer/commit/3328b81211cb2e9ae16a0b49097744171c8c71c0)
	- Autoloading subseries by time is an option break up some 4D acquisitions into individual volume, but is optional since some volumes are also acquired in time unites and should not be split.

[jcfr]

re: epsilon

If the epsilon used in DICOMUtils.getSortedImageFiles and the one introduced in the examineFile are indeed different, we should name the prefix the parameter like we have done for other epsilon (cornerEpsilon, zeroEpsilon, ...)

For context, see #7362 (comment)

cc: @lassoan @pieper

[pieper]

should name the prefix the parameter like we have done for other epsilon

Agreed!

[Punzo]

@jcfr I have applied last round of feedback!

[lassoan]

Thank you all for the reviews, this looks good to me. I just added a remark on improving comments.

[Punzo]

Thank you all for the reviews, this looks good to me. I just added a remark on improving comments.

Thanks Andras for the feedback. comments applied!

[jcfr]

Pending local testing is successful 🚀

[jcfr]

@Punzo Could you confirm that you performed local testing using the latest version ?

[jcfr]

Once this is integrated:

• Add a short discourse post describing the feature (associated with the feature¹ tag)
• Add a note to the following post cross-referencing this feature post:
  • https://discourse.slicer.org/t/facing-issue-with-using-volume-rendering/32695
  • https://discourse.slicer.org/t/dicom-file-loading-spacing/5469/4

There are probably other posts that could be updated. I used the query harden regularization²

Footnotes

1. https://discourse.slicer.org/tag/feature ↩

2. https://discourse.slicer.org/search?q=harden%20regularization%20order%3Alatest ↩

[Punzo]

@Punzo Could you confirm that you performed local testing using the latest version ?

@jcfr yes I tested locally with several DICOM datasets.

[Punzo]

Once this is integrated:

• Add a short discourse post describing the feature (associated with the feature1 tag)

• Add a note to the following post cross-referencing this feature post:

  • https://discourse.slicer.org/t/facing-issue-with-using-volume-rendering/32695
  • https://discourse.slicer.org/t/dicom-file-loading-spacing/5469/4

There are probably other posts that could be updated. I used the query harden regularization2

Footnotes

1. https://discourse.slicer.org/tag/feature ↩
2. https://discourse.slicer.org/search?q=harden%20regularization%20order%3Alatest ↩

ok!

[Punzo]

Once this is integrated:

• Add a short discourse post describing the feature (associated with the feature1 tag)

• Add a note to the following post cross-referencing this feature post:

  • https://discourse.slicer.org/t/facing-issue-with-using-volume-rendering/32695
  • https://discourse.slicer.org/t/dicom-file-loading-spacing/5469/4

There are probably other posts that could be updated. I used the query harden regularization2

Footnotes

1. https://discourse.slicer.org/tag/feature ↩
2. https://discourse.slicer.org/search?q=harden%20regularization%20order%3Alatest ↩

Done