Defining and counting pixels in an abstract area in ImageJ

Hi everyone!

I’m fairly new to imageJ and typically use inForm image analysis software. However, InForm can’t seem to do the task I currently need, and I’d love some insight on whether it can be done with imagej!

I am analyzing some 20x images of mouse adipose tissue that has been stained with picrosirius red (collagen staining) and there are several images in which the tissue does not occupy the complete image. Because of this, there are often oddly shaped areas of white background included in the images that are skewing my results. I would like to define these areas and calculate the number of pixels in them so that I can subtract them from the total. Once I can determine the number of pixels that are occupied by tissue, the goal is to express pixels positive for picred staining (I already have this number) / total pixels of tissue.

I have considered counting white pixels by making binary versions of the images, but this does not work with adipose tissue as there are large white holes everywhere where the fat deposits were located. I have also tried counting the adipocytes and normalizing the images to cell number, but some images have extensive collagen fiber staining which makes it impossible to count the adipocytes (there is no nuclear counterstain).

I would greatly appreciate any help!

Dear Tiffany,
It is a bit difficult to answer your question without seeing a sample image. However, if you can satisfactorily threshold your image to separate the tissue from the background, the Particle Analyzer (Analyze -> Analyze Particles …) should help with your problem. I would try it with both ‘Exclude on edges’ and ‘Include holes’ options ticked. That should exclude any background that ouches the edges, but include any fat deposits within your tissue (holes).
The macro version of the command would look something like:

run(“Analyze Particles…”, “display exclude clear include add”);

Hope this helps,
Volko