An automated approach for hemocytometer cell counting based on image-processing method

Abstract

Evaluating cell viability plays an important role in routine cell culture, toxicology, and drug development. Trypan blue stain is one of the simplest assays that are used to determine the number of viable cells in a cell suspension. The stained cell suspension can then be counted under a microscope using a hemocytometer. Hemocytometers have a fixed volume, making it possible to calculate the concentration of cells in the sample and measure viability. Most of the related studies are conducted using manual counting methods. However, manual counting is time-consuming and laborious. While various automated cell counting devices have been proposed, they tend to be too expensive for widespread deployment. Although low-cost automated cell counting methods presented in the literatures provide high accuracy and cost-effective alternatives, only cells in the field of view of microscope with unknown volume are counted. This will make the concentration inestimable. Accordingly, the present study proposes a low-cost but accurate method for hemocytometer cell counting using an image-processing approach. The accuracy of the proposed method is demonstrated by comparing the cell counting results obtained for human cervical carcinoma (HeLa) cells with the manual counts. It is shown that the proposed approach has a mean absolute percentage error of less than 6.26 % for samples containing a mixture of live and dead cells. Moreover, the proposed system is equipped with a simple and intuitive human machine interface such that the counting task can be performed by individuals with no prior experience in the image-processing field.