Colony counting

I’ve used your sample for colony counting for E.Coli plates. I have tried several plates and the script seems always to over-estimate the number of colonies and it includes also the bubbles on the plate. Are there any parameters that need to be changed. Or maybe the quality of the pictures it’s not good? Could you please provide some tips (I attached a sample plate with around 40 colonies counted manually)?

Thanks a lot,
Mihai (731 KB)

Hi Mihai,

It seems that the image quality is drastically different between the yeast counting example pipeline and the sample image you have here. With the yeast colony image, there is an inherent color difference that the pipeline can take advantage of to differentiate the plate from the colonies. Unfortunately, there is almost no contrast difference to speak of between the E. coli and the plate, so there is not much in the way of features to detect.

One solution is to increase the contrast by using a culture medium that will create a higher degree of contrast with respect to the colonies. Do you have a choice in the color of agar (or whatever substrate) that you are using? Alternately, perhaps adjustments in lighting may be in order to enhance the contrast through better illumination.