Automated Counting via Multicolumn Network and CytoSMART Exact FL Microscope

Neubauer chamber cell counting in microbiological culture plates is a laborious task that relies on technical expertise. As a result, efforts have been made to advance computer vision based approaches, increasing efficiency and reliability by quantitatively analyzing microorganisms and calculating their characteristics, biomass concentration and biological activity. However, the variability that still persists in these processes poses a challenge that is yet to be overcome. In this paper, a solution is proposed that adopts convolutional neural networks for automatic cell counting. The algorithm seeks to identify the characteristics of cells of various sizes using a multi-column network where there are general convolutional layers in the body of U-net. Furthermore, the solution has been implemented in the laboratory using CytoSMART Exact FL microscope images. The results show that the proposed method can handle different types of images with promising accuracy.