Something unexpected is occurring with measureobjectneighbors which may be a fault of my expectations or the program.
The module finds more neighbors than expected although the percent touching is as expected. A cell may have 0% touching but still may be identified as having >0 neighbors. I have set the module for adjacent as well as neighbor distance 1 with similar results.
I’ve attached a sample output from the module. You can see 6 nuclei have been identified. Only 2 are visually touching, but the module suggests that 5 have 2 neighbors while 1 only had a single neighbor (this is recapitulated in the output mat file). In contradiction, percent touching is correct for the 6 cells with 4 having 0% and 2 having a higher percent.
Additional clarification: Module help says that it takes into account objects that were found and discarded, but I don’t believe the percentage calculated shows objects discarded due to being on the boundary.