Issues with Image J Thresholding?

Image Analysis
,
#1

I am trying to perform colour deconvolution, but when I convert myimage to binary, I am unable to manually adjust the threshold to accurately differentiate cell bodies. As soon as I adjust the scale bar the image ether goes completely white or black. The first image is immediately after I make binary, the second when I attempt to adjust threshold.

Capture
Capture1293×845 176 KB
Capture2
Capture21280×851 46.6 KB

Has anyone had this issue before? I have tried updating my computer (mac), updating image J, using a previous version of image J, and also trying on my windows PC, to no avail. Any advice would be amazing!

#2

Hey @asorbyadams,

could you please precisely list the steps you are doing? And example image would also be good so that we can reproduce your procedure.

Cheers,
Robert

#3

Do I understand correctly that you are trying to adjust the threshold after binarising your image? If so, your image only has black and white pixels (in your case of an 8-bit file, the values are 0 or 255). If you then try to threshold between 106 and 171, as shown on your second image, you’ll get nothing…

Sincerely,

Matthieu

#4

Hi Robert, Thank you for your response!

I am currently undergoing the following with the attached image:

40xzoom microglia
40xzoom microglia2560×1440 439 KB

  1. Open image
  2. Process-Binary-Make Binary
  3. Image-Adjust-Threshold

This is the step I get to where I am unable to move the Max or Min threshold bars to accurately adjust the image threshold.I have not touched any other settings.

Thanks so much again!

#5

The problem you have is that you have binarised before thresholding, which is the wrong order. If you want to -for example- select the brown cells in your image you need to

  1. Open image
  2. Image>Color>Split Channel
  3. Select the grey channel that shows your cells best (blue or green)
  4. Threshold it appropriately and click “Apply”
  5. you now have a binary image where your cells are white and everything else is black.

Sincerely,

M.