[gasperpodobnik]

Fixes #6137 .

Description

A user can now pass a spacing parameter to surface distance metrics (Hausdorff distance, surface distance, surface dice), so that correct results can be obtained for images with non-isotropic spacing (e.g. (0.5, 0.5, 2) mm).
If spacing is a sequence, it must be of length equal to the image dimensions; if a single number, this spacing is used for all axes. If None, spacing of unity is used. Defaults to Noneso that current behaviour is preserved.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

[wyli]

thank you, could you please add a few test cases for example in this file: https://github.com/Project-MONAI/MONAI/blob/dev/tests/test_surface_distance.py

then I can edit the PR to address the mypy issues

[gasperpodobnik]

Hey, I noticed that the typical usage of the metrics classes is to simply call them (i.e. metric_class(y_pred, y)). This means that a __call__() method that is inherited from the parent IterationMetric class is used. However, this method does not support additional arguments, so there is no way to pass the spacing argument to the _compute_tensor method of the original metric_class (e.g. SurfaceDistanceMetric).
I think the best solution would be to add a spacing argument to the __call__() method of parent classes or to implement a new class (e.g. CumulativeIterationDistanceMetric) that inherits from lower-level classes and supports the handling of the spacing argument. This could then be passed as a parent class to all user-level metrics classes (i.e. SurfaceDistanceMetric, HausdorffDistanceMetric, etc.) that require the spacing argument for the correct calculation of results.
Let me know what you think is the best way of doing it :)

[wyli]

thanks @gasperpodobnik, I think we can add **kwargs to here

MONAI/monai/metrics/metric.py

Line 321 in 600621d

self, y_pred: TensorOrList, y: TensorOrList | None = None

and

MONAI/monai/metrics/metric.py

Line 52 in 600621d

self, y_pred: TensorOrList, y: TensorOrList | None = None

for additional parameters, could you please help revise if you agree?

[gasperpodobnik]

sure :)
Do you maybe know which tool was used to get the reference measurements of Hausdorff distance, surface distance, etc. for testing?

[wyli]

/build

[wyli]

many thanks, it looks good to me.