Help with CellProfiler 4.0.4 and cell segmentation

Sample image and/or code

Background

I need help with CP pipeline to identify Red Nuclei of cells in visual cortex in two photon calcium imaging.
(Green color is for activated cells. Don’t need to identify these.)

3 separate individual PNG image files, each correspond to a time averaged image capture.
This is not Z-stack.

Analysis goals

I need help with CP pipeline to identify Red Nuclei of cells in visual cortex in two photon calcium imaging
and count the number of Red nuclei in each of 3 PNG images.

I also need to save the 3 output image files, with Red nuclei circled and marked by CP.

I am new to CP, so any help appreciated!

Challenges

  • I am new to CP.
  • I downloaded CP 4.0.4, ran example file “ExampleFly” pipeline and it works well.
  • No
  • CP 4.0.4,

Clarification: Need to identify and count Red channel only, which is nucleus of cell.

Also need to save the perimeter/boundary of nucleus/cell marked by CP, as binary mask or something similar, so I can post-process it. Any help appreciated!

Hi @ramana,

Have you checked out some of our video tutorials that might help with this?

FOV3_donut_gray_0_4_outline.tif (41.0 KB)

Thank you! I am able to create my first pipeline from your video tutorial!

I have another image file, atatched above, which has donut/ring shaped cells and CP misses some of them. I specified “Fill holes in objects after declumping and thresholding” but it does not help.

Any pointers appreciated!

correct image attached.

Here are a couple ideas:

  • Apply a circular Hough filter. This would enhance ring, or circular, objects. Back in CP 2 we had such a filter, but I don’t think it’s in the current version, CellProfiler 4.0, so let’s not worry about this right now
  • Use EnhanceOrSuppressFeatures to enhance ‘dark holes’. Dark holes isn’t exactly the same as ring-shaped objects, but I think it can actually work pretty well for you. I created a quick CP 4.0.7 version of a pipeline to do this for your image.
    • Be sure to read the comments in the pipeline for each module!

Good luck,

David
Akhila_Raman_Holes_pipeline_CP4.0.7.cppipe (7.1 KB)

Thanks, David!

I was wondering if CP has an option to display only specific output plots like
“IdentifyPrimaryObjects #7” and “OverlayOutlines #8”?
right now CP displays so many plots , it clutters screen.

Also, does CP have " background subtraction " for noisy image below?

Yes, the little eyeball next to each module.

The Reduce Noise module will reduce random (“shot”) pixel noise, however I don’t see that in this particular image. I see what I might call “background” staining, but even that is a guess at the origin of the lighter staining. If you are looking to ID the cells with holes in the middle I would simply use IdentifyPrimaryObjects and see if you can get the thresholds set to do your filtering at the same time as segmentation (along with Dark Holes in EnhanceOrSuppressFeatures still).