Congratulations on getting CellProfiler 3.0 released. I have been testing out the new 3D segmentation functionality over the last couple of weeks. I think I hit a roadblock. so I wanted to get some feedback. I am finding it hard to declump closely spaced nuclei. The best option that I have come up with are to use an active contour module and a downstream watershed module set to “generate from distance” and a connectivity of 26. I find that this tends to declump the nuclei a little better than the active contour module alone. However, it appears to drop nuclei in the watershed step. Has anyone encountered this bug during testing? I would like to hear if there are any other potential declumping options that could do a better job. Let me know if you would like me to post a pipeline for testing.
Feature Request - SplitOrMergeObjects for 3D #3363