kevinjohncutler
commented
1 year ago @nikamarand sorry about that, I forgot to update the ncolor package (that return_n is something I implemented a few weeks ago). It is now on the ncolor github: https://github.com/kevinjohncutler/ncolor
Hi Kevin, I ran into an issue when running through your example notebooks for mono channel bacteria/multi channel bacteria. I installed omnipose and cellpose-omni directly from github as directed from your readme file in omnipose, so I have the most up to date packages. When plotting your example images, I received a kwargs argument when plotting segmentation with ncolor. The argument was for return_n -> this is from your plot.py script what was pulled from cellpose-omni package (line 141): m,n = ncolor.label(maski,max_depth=20, return_n=True) I changed the code slightly and made n=4 based on your color.label def and changed the above equation to m = ncolor.label(maski, max_depth=20) This at least lets me run your jupyter notebook without issue. Just not sure if there is a bug somewhere in your most up to date code for ncolor or if something didn't get updated to work with omnipose? In terminal using my own images, something I believe to do with ncolor as well, I get an output image that is black instead of a mask of my cells colored by ncolor, which then doesn’t produce subsequent outlines. Not sure if these two issues are related though. Any help/clarification on this issue would be much appreciated.