Inconsistent grayscale inversion


I am having this issue where the same version of ImageJ - I’m running v.1.52p33 to make getValue(“rotation.angle”) to work - running on two different computers gives the exact opposite response to thresholding and analyze particles (possibly inverting what it considers to be light and dark). Has anyone had this issue before? Below is the an example image.

A.tif (1.1 MB)

I run the following steps on this image -

run("Auto Threshold", "method=Otsu");
run("Invert LUT");
run("Analyze Particles...", "  show=Masks clear summarize");

One PC gives me 14.72% under %Area, and another PC gives me 85.28%.

The image after thresholding in each PC are the opposite of each other (in terms of what’s light and what’s dark). Has anyone had any experience? For now my fix is that as I know that my %Area is going to be below a certain number, I’ll adjust my image grayscale accordingly. But this is not really the right way of dealing with this issue.

Any guidance would be of great help.

Thank you.
Kind regards,
Somsubhro Chaudhuri

Two suggestions:

  1. Check the "Black Background "option in Process>Binary>Options. It might be set differently in the two machines.
  2. Avoid “inverted LUTs” You are not inverting anything, just the display. Use Edit>Invert instead to do a proper inversion.

Point 1 was spot on. Thank you @gabriel.

Thank you for point 2. I did that just to get around this issue. Now I won’t have to anymore :slight_smile: