Can a RandomizableTransform depend on a previous Transform?

[jonieva]

Hi there!

I was wondering if there is a way to make a RandomizableTransform depend on a previous RandomizableTransform during a training pipeline.

During training, I would like to do something like:

• Apply RandomizableTransformA
• Apply RandomizableTransformB
• If RandomizableTransformB was not applied to this image, apply RandomizableTransformC.

I know I can override the randomize method in my transform, but I'm not sure how could I know if a previous transform in the Compose pipeline was applied.

A workaround could be to implement a "composed Transform" RandomizableTransform_B_or_C, but I was wondering if someone has a better alternative.

[wyli]

thanks for the discussion, does the OneOf transform as an extension of Compose solve this problem?

MONAI/monai/transforms/compose.py

Lines 385 to 388 in f140e06

	class OneOf(Compose):
	"""
	``OneOf`` provides the ability to randomly choose one transform out of a
	list of callables with pre-defined probabilities for each.

[jonieva]

I think it does! Or at least it's definitely much cleaner :-)

Thank you very much @wyli