Doing consistent image analysis


I need help to analyse these two files of images !

I am trying to do ratiometric measurments means( Magenta/ Red images )
The workflow is:

  1. Images >>> Stacks >>> Stack to images
  2. Images calculator then I divide 2/3 image by 3/3 image
  3. Substract the control mean intensity from the resulted image
  4. Double
  5. Threshould
  6. Edit >>> selection >>>creat selection of thresholded image
  7. Choose the original result image >>> Edit >>>> restore selection
  8. Analyse >>>> set measurements >> chose the result image
  9. Measure >>> Mean intensity
    I am intresting to see the change of ratiometric measurements in response to oxygen concentration.
    Each resulted image has a britghtness and contrast, is that affect the mean intensity?
    Should I keep the maximum and minimum brigtness the same to get right intensity?

Hi Reem,
Adjusting the Brightness and Contrast values only affects the actual pixel values if you click ‘Apply’ on the Brightness/Contrast dialog, otherwise they just affect the way the image is displayed on the screen.
I always thought that there should be some form of warning that clicking ‘Apply’ does irreversibly change the pixel values as I have seen many students accidently changing their images.
Hope this answers your question.


I agree with Volko that the measured intensity values will not change if you don’t apply the contrast. If you (and only in that case) want to present your results with example images, however, you should make sure, that the B&C between conditions is identical.

I do however have a question about step 3 in your workflow. What exactly are you subtracting there?

  • Is it the mean intensity of one of the original channels (beacause I think that might produce misleading data after dividing the two channels).
  • Or is it the control group mean unrelated to each individual image, so some kind of normalizing step of treatment to control? That might be easier to do outside of ImageJ and it would allow you to do the exact same analysis steps for both control and treatment images.

A final suggestion: It seems like you are doing this analysis manually for each step. If you start the macro recorder for one of your images, you probably will end up saving a lot of time and have an even more reproducible analysis :slight_smile: