01 January 1970 2 3K Report

I am really keen to have your opinions on this fluorescence image analysis with Fiji/ImageJ. I am trying to measure the mean value of GFP in the polygonal ROI labelled in Original Image, which will then allow me to compare the expression of the GFP-labelled protein in the same region of the tissue between two groups.

I am thinking about 3 different ways:

  • Measure the mean value (1) within this ROI, then measure the background mean value (2) by using a set of small ROIs placed in those dark areas within the above polygonal ROI. Then subtract (2) from (1). Easy to achieve, and seems reasonably logical.
  • Threshold the image using range of 0 to value (2) as measured above to get the Thresholded Image. The white areas then have value of 0 and black areas 255. Then, if possible, I could “map” the black areas to my original image to measure the mean value from the “proper” protein expressed area (let’s assume that the autofluorescence is comparable between 2 groups). Now comes the question – how do I create ROIs from the thresholded image to cover all black areas (magic wand only highlight individual white areas)? Or is there a way to do the “mapping”?
  • Draw a lots of “islands” (i.e. ROIs) to cover all the GFP expressed areas in the original image, measure the mean value (3), then subtract (2) from (3). I feel this method is quite subjective, likely to create great variabilities between images.
  • ImageJ/Fiji experts, please share your thoughts and different ideas from what I listed above. Many thanks!

    More Lan Zhu's questions See All
    Similar questions and discussions