Thresholding Analysis w/Color deconvolution


I am analyzing nerve sections stained with trichrome in Fiji. After running color deconvolution, I am thresholding the resulting images but have some basic questions that I haven’t been able to find a straightforward answer for.

After thresholding an image, the regions of interest are highlighted in red. By then presssing “Apply”, the red highlighted areas become black. I have Measured both the red and black images for % area, mean gray value, and integrated density. The % area does not change for the red and black images, but the mean gray value does. The mean gray value and integrated density vary for the two images, and I’m not sure which image to analyze going forward, and what the differences between the two images mean. Additionally, the min/max threshold for the black image is always 255, while it varies for the red image. I’m assuming the black image is made binary, while the red image is not, but I’m not sure what this means for the integrated density measurement and which image would be best to analyze.

Thanks for any help you can provide.

From the sounds of it, the red image is your original image. The red in FIJI just shows you where the threshold is exceeded. Once you hit apply, it creates a binary mask that is 0 or 255. The mean grey value of the mask image would, I guess, be proportional to the masked area.


Maybe a few example images would help.
Generally, when you press the Apply button, the red highlighted areas become white. Pay attention to the options (black background…)
% area, mean gray value and integrated density should be linked.
For example, if you have 70% of white pixels, mean gray value should be 255*.70 = 178.5. integrated density should be 178.5*number of pixels.
If this is not the case, it’s probably that you don’t measure the same thing on both images.