Nerve quantifications : axons and myelin thickness

Hello everyone!

I’m trying to quantify the size of axons and myelin thickness in cross section of sciatic nerves.
Unfortunately, the color of the axons ( the part inside the blue circles) is the same as those parts that are not to be counted and when I’m trying to separate them not always I get a nice separation.
I have used Ilastik to segment the image and then used the probability masks derived from there in CellProfiler. I cannot get the pipeline to identify objects correctly, even if I try to exclude by size or circularity.
my only other alternative right now is using Fiji and measure every single of them and is taking forever!
Any suggestion?
I have attached the original image, the segmented one from ilastik and my pipeline.
Any help will be appreciated!

SciaticNerve.cpproj (164.4 KB)