in the context of a recent PR by Simon (thanks for working on it) to add an implementation of the Watershed algorithm, I have asked in the
imagej-ops Gitter channel how to handle the scenario where you’d want to run the op only on specific parts of an image. In this particular case, the implementation features an optional
mask parameter that is used to determine which input samples are taken into account during the computation.
I was wondering if this is a more general concept and if you could come up with scenarios where such functionality might also be useful?