CP 3.0 seeded watershed bug



I think I might have found a bug in the RemoveHoles module when there is more than one nuclei seed per cell. I am generating nuclei seeds in Matlab which are passed into the 3D seeded watershed segmentation step to generate the cell body segmentations. I setup the pipeline to filter the small objects generated from the extra nuclei seeds which is working ok. The problem comes when I try to remove the holes generated by the filtering step. As you can see in the image I posted below that all the cells get removed in the output stack. Any ideas for a workaround?

Github issue: “RemoveHoles module removes all cells if there are more than one nuclei seed per cell” #3369