Viability module thresholding issues: Noise identified as cell, cell not identified.

[lindsayspen]

Hi all, 

I'm having some issues with using the viability module to count cells in a counting chamber.

The thresholding seems to work fine except that many cells which are adequately outlined are not identified (no number next to them). If I lower the threshold to try and make them identified, noise comes in (bottom right corner) and is immediately identified as cells and the original unidentified cells remain unidentified. 

Here are images of the problem. Threshold of 40 on the left misses many cells in the centre of the image. Threshold of 36 on the right still misses cells in the middle but falsely identifies lots of noise in the corner. The problem gets worse with lower threshold.

Is there a way around this other than better image quality?

Help is appreciated, Thanks,
Lindsay

[Carlos Alquezar]

Hi Lindsay,

maybe I am wrong, but it seems that your upper bound of cell size is a bit small. In the viability settings, try to increase the maximum cell size. Also, check that you are working with the properly microns/pixel ratio.

I hope it helps you and please, let me know if it works.

Best,

Carlos

[lindsayspen]

Hi Carlos, 

Yes you were right thank you, a higher upper size limit produced much better results.
I have been using the freehand select tool in combination with set scale and measure to estimate cell size. But the results don't seem to correspond to what OpenCASA measures. Is there a better way or should I just input a wider size range?

Thanks again, 
Lindsay

[Carlos Alquezar]

With which species do you use to work? I would recommend you to look for the specific size for your species in the literature. Also, you could use the morphometry module (setting the corresponding microns/pixel ratio properly) to estimate the cell size.

Best,

Carlos