I am calculating Fluorescence Anisotropy of live cells expressing EGFP. Due to light scattering and presence of other molecules, the background intensity values of my images are pretty high (~1500 counts for a 16-bit image). Now, in this case, I am considering those regions in the field of view having no cells as background and subtracting that number. Is this the proper way for background correction?

Also, I have found that the same imaging media (DMEM with phenol red) without any cultured cells give a count of ~800 at the same exposure and lamp power. So, is that the proper background?

Again, the dark current of my camera gives a count of ~150, as I found by imaging just water.

Subtracting different backgrounds give different values of Anisotropy. I am confused as to what value should I use. Or do I even correct for background at all, because most of the light I observe comes from the point spread function for the objective-fluorophore.