A Method of Automatic Cell Counting Based on Microscopic Image

Abstract

This paper introduces an efficient automatic cell counter based on microscopic image. The counter separates the cell and background by histogram dual-threshold, fills the cell by Floodfill, and detects the cell using blob analysis. Then, using K-means clustering method segments the adherent cells. The highlight advantage of our proposed method is that it has good adaptability for different kinds of cells and strong robustness to light intensity. Experimental results show that the counter can quickly and correctly detect the number of cells and have good stability, regardless of the cell morphology, cell type and cell size. The method has been successfully applied to the microscopic image processing system.