Correcting for illumination across image sets

I’ve been working on identifying condensed spheres of DNA in the nuclei of a particular mutant line of cells. I’ve been able to make a pipeline (attached) that identifies nuclei and then identifies the “gumballs” of DNA, as we call them, within the nucleus. When there is just the nucleus is identifies that as a single “gumball”. The problem is that it identifies the single gumball just fine in one control treatment (Smc2 in the attached pic) and doesn’t identify the single gumball in another control treatment (SK_control, attached). Is there a way to normalize the illumination across all the treatments / image sets? Or is it ok to adjust the threshold for one treatment (in this case, SK_Control) and use another threshold for the other treatments? I also attached a “gumball” treatment (SkpA) in case that helps.
nuclei_and_gumball_or_no_gumball.cp (8.43 KB)

Hi Scott,

In the examples you uploaded, it’s a bit hard for me to tell that the identification is proceeding correctly for the control images since it still identifies a number of blobs (presumably non-gumballs) in the nuclei. I presume you’ll want to do some post-identification filtering of some kind?

Even so, to answer your question, yes, it is possible to include illumination correction as pre-processing step using the CorrectIllumination_Calculate and _Apply modules. Furthermore, this can be done on a per-object basis by using the cropped image that you’ve created. I’m attaching an example pipeline which describes what I mean; using the result as input into IdentifyPrimary, it seems to detect just a few blobs for the control images and more, larger blobs in the gumball-postive image.

2011_08_10.cp (10.3 KB)