Measuring the sellsize - identify the cells


I’m new to ImageJ and don’t really know how to solve my problem.
I want to measure the cell sizes/areas and to count the cells, but because of the low contrast between the cells and the base area and the light reflection of the cells the threshhold can’t distinguish between cell areas and the base area. The function “analyze particles” then just creates me some big and wrong ellipses because the threshholded picture isn’t clear enough.
I’ve marked exemplary some of the cells as ellipses.

I would be very grateful if you could help.


If you have several images, you could use ilastik to train your segmentation and then save those segmented images and use them in Fiji or CellProfiler.

1 Like

Hi @ki11erviet,

You can also try in the CellProfiler directly. I tried with your image. I have tried with a very basic pipeline with EnhanceSuppress module & IdentifyPrimary object module. It looks to be good start. This is what I got,

It is identifying most of the regions. But I am sure you can start here & add modules to reduce the noise.
To remove the light reflection, you could try illumination correction module. You can also refer to cellprofiler tutorials & examples
Also attaching the pipeline heremeasure_cellSize.cpproj (659.6 KB)
Hope this helps.

Fujifilm Wako Automation (Consultant)
For CellProfiler training or optimised pipeline write to,

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS

Join us at,