Help: I need to count the number of cell positive to the membrane stain

Hello, I need to count the number of cell positive to the membrane stain. Here an example of image i would like to work on:
The first thing I did was to separate the channels and try to improve the image (green channel) in order to have defined contours and be able to use analyze particles. My problem is that I am not able to optimize the image.

could you please highlight some positive and some negative cells?

Dear Sara1,

You can try to use a marker controlled watershed by detecting the nuclei and use them as markers and use the green channel for separating the different cells.

The MorphoLibJ library offeres a number of very good algorithms for that:

To generate the markers apply a gaussian filter:
Process > Filters > Gaussian Blur…
With a rather large sigma, the size of your nuclei
Then you can detect each nucleus using the Maximum Detection:
Process > Find Maxima…

Once you have each cell segmented you can try to measure.
But here lies the true challenge of your images. Since the signal seems to be at the borders you will have a tough time to differentiate positive from negative cells…
especially since the delineation of each cell depends on the signal you want to measure and this is probably gone with negative cells…

1 Like