Incorrect number of neighbors near image border

I’m noticing a problem with the number of neighbors for objects detected next to the ones that were discarded at the image’s border. Although the objects on the border were discarded, their existence should still be maintained so neighboring objects are measured correctly. That’s what the CellProfiler manual says, but our data says different. For example, a cell that should have 6 neighbors (2 are border cells) has a count of 4 neighbors in the output data. Do I have incorrect settings in my pipeline, or is this a bug in the software?
pipeline2.cp (11.6 KB)


I believe this is a bug (one that I thought we corrected previously?), and I created a bug report here: We’ll post here when it’s fixed.


This has been fixed. You can try a “bleeding-edge” version here: … TrunkBuild

Note that 2.0 more closely matches the user interface and functionality of the current release.


This issue has been resolved in CellProfiler 2.1 and later, new releases of which can now be downloaded from