Why lead to such result?

cellprofiler

#1

When I run my modified YeastColoniesPIPE on my attached gray image, I don’t konw why the number of colonies is 0.
While the number of colonies is easily counted by manual.
Is there a minimum limitation in colonies number?

My modified pipeline is as follows.
LoadImages
CorrectIllumination_Calculation
Correctillumination_Apply
LoadSingleImage
ApplyThreshold
Align
Crop
IdentifyPrimAutomatic
MeasureObjectIntensity
OverlayOutlines
SaveImages



#2

No, there are no arbitrary limits on the numbers of objects imposed. Likely your IdentifyPrimAutomatic settings need to be altered (but it could be the illumination correction, or the cropping, etc.). You’ll need to display each module’s windows and carefully look through each one to discern what it’s doing. If it’s not clear to you, then upload your pipeline and we’ll take a look.

David


#3

Hi,
My pipeline is as follows.
LoadImages
CorrectIllumination_Calculation
Correctillumination_Apply
LoadSingleImage
ApplyThreshold
Align
Crop
IdentifyPrimAutomatic
MeasureObjectIntensity
OverlayOutlines
SaveImages

Maybe, there are some errors in main modules. Please correct it for me ,thank you.
Meanwhile, tell me the settings of the corrected modules.


#4

Hi,

Could you please upload your pipeline .mat file, rather than list the modules? It would be much easier for us to work with. When you post a reply, click the “Upload attachment” tab below edit window.

Thanks,
David


#5

I am sorry. Thanks a lot. I have uploaded .mat file.

By the way, in Cellprofiler, is there a module that can merges two or more images so that we can see whether there are some overlapping spots ?
ModifiedGrayYeast0819PIPE.mat (1.63 KB)


#6

Hi,

To answer your question about merging: Try using the GrayToColor module. You can assign a color channel for each image for up to three grayscale images and then merge them into a color image.
-Mark


#7

And, if you find that the alignment is poor between the color channels, you can use the Align module to align the individual channels prior to using GrayToColor.

And note, sometimes your image looks grayscale but it’s actually been saved as a color image format. In that case, you would need a ColorToGray module to convert each grayscale-looking image to a real grayscale image prior to making a color image as Mark suggests.

Anne


#8

Hi,

Here’s a pipeline that works. I played around rescaling the intensity and enhancing the spots, because CellProfiler was missing them because they were so dim. Did you by any chance photoshop this image first? Sometimes images look very dark, but the camera they were captured on may have been 12bit, saving them as 16bit. If you photoshopped it at all so you could see the image, sometimes that messes up the scale of the intensity. That would be my guess, since rescaling worked.

~kate
ModifiedGrayYeast0819PIPE.mat (1.88 KB)


#9

Hi, I have runned the pipeline you upload. But there is a error.
It points out that image processing was canceled in the IdentifyPrimAutomatic module because Laplace values are invalid.


#10

Oops! This is my fault. I created the Pipeline in the developer’s version. I will re-create the pipeline in ver 5122, but in the meantime if you try using the new version released on the website yesterday(5811), it will probably work.


#11

Here you go, I made this pipeline in version 5122, so you shouldn’t get an error.

~kate
yeastPIPE.mat (1.79 KB)


#12

I have successfully runed that pipeline. Thanks a lot!