Exclude saturated pixels in measurement

Hi, I want to use cellprofiler to measure fluorescence of 2 different channels (CFP and FRET) in a time course experiment. The fluorescent cells are growing in a array format and each spot represents a clump of fluorescent cells. For each spot I would like to measure the change of fluorescence in a time course experiment for only the fluorescent cells. I have already made a pipeline that enables me to measure the intensities of separate spots for both channels. However, many spots contain some saturated pixels which I would like to exclude from my measurement. Is there a way to do this?
Furthermore, I would like to know whether it is possible to read Leica .lif files in cellprofiler?
Thanks in advance!

20140423 Cellprofiler test.cpproj (75 KB)

Re: saturated region exclusion - Our typical strategy is to estimate the intensity that deem as saturated using MeasureObjectIntensity along with DisplayDataOnImage for visualization if you need it. Once you know what intensity value sufices as a cutoff, you can use FilterObjects to remove these objects from consideration.

Re: Leica files - Yes, CP 2.1 should be able to read them. If they are stacks rather than single images, you may need to have CP extract the image information from the file header so that it knows how to handle them. If this is the case, check out this tutorial for details: cellprofiler.org/tutorials/L … movies.pdf


Yes, the simplest way to remove the saturated pixels is to threshold them out so that their intensity is zero (or whatever value you deem, actually). You can use ApplyThreshold to do this, as I have in the pipeline attached. You need to choose the saturation threshold – I chose 0.99 arbitrarily; note that CP scales intensities from 0 to 1 internally.
Alternatively, if you want to do this on a per-object basis, you could MeasureObjectIntensity as you already are, then use FilterObjects to filter based on say, Max- or Mean-Intensity, and then use a subsequent MeasureObjectIntensity on the filtered output. But I suggest the former, pixel-based method.

As for .lif, CellProfiler utilizes the image library Bio-Formats to read all sorts of image formats, and it does apparently read .lif (openmicroscopy.org/site/supp … rmats.html). If you are having a problem reading these or setting up the Images module, post an example file and we’ll take a look.

20140423 Cellprofiler test DL.cpproj (422 KB)

oops, sorry Mark, we double-teamed this one!

Hi guys,

Thanks for the quick reply for the both of you! I will have a look at both your suggestions.

Best regards, Maurice