help!! with neuron soma counting w high background noise

[Hihan]

hello,

I am attempting to pick up neuronal somas (stained with an immunohistochemical protocol)  in a sectioned brain. This neuron (kisspeptin) has high fiber density in the background so the images have very dark background which confuses Image J when I try to set a threshold. The neuron somas are also ununiformly spread out and not perfectly concentric. I know it would be great to do a nuclear staining, but I don't have the time or the resources to carry on with that method.

Is there anyway that I can computationally separate somas from the high density fibers?



I am also a noob at Image J.. is there a tutorial that I can take that addresses high background noise treatment.?