Adipocyte pipeline- Identification problem

Dear sir, madam,

I am a masters student from Wageningen University (the Netherlands) and using CellProfiler with Adipocyte pipeline to measure adipocyte size for our research on the effect of early life nutrition on later life gonadal WAT of female mice.
CellProfiler is really helpful in our research and I am pleased to be using the program.

However, I encountered some problems with three of my samples which I did not encounter in all other 15 samples. CellProfiler seems to be (partly) unable to identify adipocytes in the pictures which all were stained in the same way with H&E.

Below you can find images of a picture which could not be identified and one which could. It contains a lot of broken adipocytes but this does not seem to be the problem as the pictures that could be identified also contained broken parts.

I used the standard settings of the pipeline but changed the min typical size to 10 and switched ’ method to distinguish clumed objects’ (in IdentifyPrimaryObjects) to None as this seemed to work better in all other samples.

Did you ever encounter the same problems and what can be done to prevent this?

Thank you in advance.



It’s hard to say without actually having the pipeline (see here for the information we recommend including for troubleshooting), but by eye it looks like the cases where it fails have thinner cell membranes delineating them than others.

Run IdentifyPrimaryObjects in test mode with output visible (the little “eye” open) on an image that doesn’t segment well- are those areas just not being picked out as objects at all or are they being picked up as one large object and then not properly segmented because the borders are too thin? This will give you an idea of whether you should start by adjusting the threshold used to identify the cells or whether you need to adopt some sort of declumping strategy after all.

Good luck!


Thank you for your advice.
I am sorry for not attaching the pipeline. However, with your help I found the solution to the problem. I changed ‘assign pixels in the middle intensity class to the foreground’ instead of background and now it works in all of my pictures.

Thanks again for your quick reply!