I think some of the background problem stems from how you are saving your color images. I cannot tell if the background staining is another wavelength/stain different from the DNA stain, or whether it is actually part of the DNA stain. If the former, then excluding this other stain from the analysis should help things; if the latter, then some other pre-processing might be needed. But as it is, the image appears to be blue objects overlaid on a gray background. Since a gray image saved as color introduces blue components where there were none before, I cannot separate the two completely in these images.
- You should save each individual wavelength/channel as a separate image.
- This image should be grayscale, rather than pseudo-colored. Using color may be nice for visualization, but it doesn’t help the analysis.
I also notice that the blue channel is saturated (i.e, maxed out in intensity in some regions). If you are able to set the exposure to a lower level, it should aid the ability to identify the nuclei.