Pixel classification of phloxine B stained yeast cells

I have been trying to use Ilastik pixel classification to count phloxine B stained yeast cells in green channel images. My problem is different intensity of green fluorescence for different yeast strains. Images were taken by fluorescent microscope, at same exposure time, but some of the images show weak fluorescent signal while others has a very strong ones, even at the same strains. Other problem is that in raw format some images are pretty dark (short exposure time). I tried changing min/max display value in all images and it worked pretty well for images with strong fluorescent signal but for images with weak fluorescent signal it turn out pretty bad. Is there a way to batch process those images so they can have similiar fluorescent signal or at least sort them by signal strenght so I can use different pixel classification projects to count my cells.

Im pretty new to image processing field so any advice would be great!

Thank you in advance!

Hi @Tugdo,

welcome to the forum!

Just a quick question, what workflow are you using in ilastik? And do I read that right that you are only interested in counts of cells?


Im using basic Pixel Classification. I just choose features, and train Ilastik on choosen images. Yes, Im interested in counting cells, but only those which are seen in green fluorescent channel. As I mentioned before, my biggest problem is a number of images (around 400) and different intensity of green fluorescence.