Confocal serial sections (detection and quantification)

I have made stacks of multi channel confocal Z- serial section images of
immunohistochemical localization of a specific type of protein found on the
intima of the thoracic aorta of rats. I have also merge the images obtained
by each channel into a single image. There are a bunch of stuff that I need
to do, and I have decided to use imagej to conduct them. Firstly, I need to
be able to detect the outline of the intima since it is very thin compared
to the media and adventitia, they all make up the wall of the aorta. I have
been looking at a technique called segmentation using snakes but I am not
sure fully how that works. Secondly, I need to quantify the fluorescent
intensity of each Z section of the stack taking into consideration that the
three dimensional fluorescent staining of the protein is not symmetric about
the z -axis. Since I am dealing in effect with hundreds of images, I need to
have this process (detection and quantification) as automated as possible.

Jimmy