Using a reference region to normalize stain intensity

I have DAB stained slides that I plan to use thresholding classifications for analysis but, the overall stain intensity of tissue differs from one image to another (example given below). Therefore if I apply the same thresholds to both images the analysis results are not comparable. Is there any way to make these images comparable? For example is there any way to select a region within a tissue section that can be used to normalize the intensity of other images? I am open to any suggestions as re-staining tissue is not a viable option for us. Thank you so much in advance for any help and/or suggestions on how to fix this.

If you aren’t correcting for background, I don’t think there is any legit way to do this. You don’t have any control on each slide or anything to tell you whether darker stain is due to the sample preparation or the staining actually being darker in that sample.

QuPath also doesn’t really do much image adjustment (though you can export new images with more complicated scripts), so you will mostly be restricted to adjusting the color vectors/background, or adjusting your thresholds per sample.

Thank you so much, I had a feeling that nothing could be done but wanted to double check. Happy almost New Year!

1 Like