Image Threshold

Hi All,

I am facing issues with setting the threshold values for images. In the attached image you can see that not all of the GFP is captured. But if I reduce the threshold value to include the projections, then the boundary for the other cells also is increased capturing void area. Is there any correction or step that can be included to correctly capture the approximate boundary of the cell?

Included the image below to highlight the clump of cells as well as the projections that haven’t been captured.

Please let me know if anyone has similar issues and how do you deal with it.

Thanks in Advance!

B10_01_1_3_DAPI_001.tiff (3.2 MB)

Maybe doing a background subtraction will give you a better segmentation results.
Here is are discussions about background subtraction in Cellprofiler:

Thankyou. I’ll check it out and see if it works!

Brst,
KD

Hi,
Did you try the background subtraction?? Hope you are segmenting your cells as a secondary objects, then try slightly different Thresholding Strategy (Propagation). You could slightly play with Threshold correction factor as well. We could help you better if you could share your sample image/pipeline.

Regards,

Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

Join us at, https://www.slas2020.org

Hello Laxmi,

Thank you for checking.

Yes, I tried the background subtraction but I have been facing an error message:

Please find attached the pipeline.

Best,
KanupriyaMicroglia_Morphology_Pipeline_10X_KD_Experiment 2New.cpproj (1.4 MB)

Hi @KanuDaga,
@bcimini

In your ImageMath module, give your image first & then the measurement, it works (I tried with dummy image). Like this,

To include your projection you could try use “EnhanceandSuppressfeatures” before secondary object segmentation. It improves your segmentation. Again with dummy images from yours,


Here is your modified pipeline.
Microglia_Morphology_Pipeline_10X_KD_Experiment 2New1.cpproj (1.1 MB)

Hope this helps.
Regards,

Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

Join us at, https://www.slas2020.org

1 Like

Hi Laxmi,

Thank you so much for the modified profile! It is working well with my image set. In my images though, there is too much reduction in background color for the secondary object specially, so much so that you can’t identify if the threshold values are appropriate or not. Is there a way to reduce the background yet keep the color for the primary and secondary objects? Please let me know.

Thank you once again for your support! Also, Happy New Year :slight_smile:

Best,
Kanupriya Daga

Hi @KanuDaga,

Thanks for the wishes & wishing you the same!!!
I am not sure what/which do you mean by background color. Nevertheless, in case you want to subtract the BG before segmenting your primary or secondary object, its worth exploring the link that is shared by @schmiedc in this thread above. Sharing the screenshot of the problem would make support us to help you better.

Lakshmi
Fujifilm Wako Automation (Consultant)
www.wakoautomation.com
For CellProfiler training or optimised pipeline write to,
lakshmi.balasubramanian.contractor@fujifilm.com

Read more on our site.
Yokogawa CV8000 - The Ultimate in Confocal HCS
https://www.wakoautomation.com/products/yokogawa-high-content-imaging

Join us at, https://www.slas2020.org