Our cell counting algorithm is a deep learning based algorithm. This means we have input a lot of cell counting images into the algorithm to learn it how to detect alive and dead cells.
The benefits of a deep learning algorithm in comparison with the classical image analysis algorithms is that you can teach it to detect more than just alive and dead cells. This helps in preventing the algorithm from counting debris and grid lines, which in some cases is inevitable. But this also made it possible to improve over time.
This is the reason that we currently have our new version 3 algorithm, which is able to detect even smaller cells down to 5 µm.
Currently we have two algorithms available, the original and new. This is to make sure we do not interfere with on going research.
The new version also comes with a multicount option, you can now take multiple pictures of the same sample. This increases the sampling area and results in an improved accuracy of your cell counts.