[Feature] nnUNet-style Gaussian Noise and Blur

[blueyo0]

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

implement nnUNet-style Gaussian Noise and Blur

Modification

Please briefly describe what modification is made in this PR.

BC-breaking (Optional)

Does the modification introduce changes that break the backward-compatibility of the downstream repos?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.

Use cases (Optional)

If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.

Checklist

1. Pre-commit or other linting tools are used to fix the potential lint issues.
2. The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
3. If the modification has potential influence on downstream projects, this PR should be tested with downstream projects, like MMDet or MMDet3D.
4. The documentation has been modified accordingly, like docstring or example tutorials.

[codecov]

Codecov Report

Base: 83.57% // Head: 83.61% // Increases project coverage by +0.04% 🎉

Coverage data is based on head (976c460) compared to base (6eb1a95).
Patch coverage: 89.06% of modified lines in pull request are covered.

Additional details and impacted files

@@             Coverage Diff             @@
##           dev-1.x    #2373      +/-   ##
===========================================
+ Coverage    83.57%   83.61%   +0.04%     
===========================================
  Files          144      144              
  Lines         8238     8302      +64     
  Branches      1226     1237      +11     
===========================================
+ Hits          6885     6942      +57     
- Misses        1160     1161       +1     
- Partials       193      199       +6     

Flag	Coverage Δ
unittests	83.61% <89.06%> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
mmseg/datasets/__init__.py	100.00% <ø> (ø)
mmseg/datasets/transforms/__init__.py	100.00% <ø> (ø)
mmseg/datasets/transforms/transforms.py	94.56% <89.06%> (-0.68%)	⬇️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[xiexinch]

Hi @blueyo0,
Thanks for your contribution, please fix the lint error. https://github.com/open-mmlab/mmsegmentation/blob/master/.github/CONTRIBUTING.md

[MeowZheng]

please add some ut for BioMedicalGaussianBlur

[MeowZheng]

From this implementation, only when self.different_sigma_per_axis is True, it will generate different sigma for each channels along Z axis?

However, different_sigma_per_axis is not working for the function described in docstring, it works as different_sigma_per_channel

different_sigma_per_axis (bool): whether to use different
            sigma for axis X and Y of the image. Default to True.

[blueyo0]

Here are some mistakes in this docstring. Actually, different_sigma_per_axis controls axis X, Y, Z, and different_sigma_per_channel controls axis N. The docstring is corrected now.

[MeowZheng]

There is a little question at L1657-1659, If different_sigma_per_axis is True, it will generate different sigma along Z? not X Y Z?

[MeowZheng]

sorry I got it, it generates sigma for data_sample.shape[1:], I just ignore :

[MeowZheng]

LGTM