Attempting to port QuPath's default cell detection to Python.
The sample image is extracted from the famous CMU-1.svs, part of the OpenSlide freely distributable data (https://openslide.org).
It doesn't perfectly replicate QuPath's method (and the watershed splitting is questionable) but it hopefully gives the general idea.