A Method of Cell Counting Based on Computer Vision

For a long time, making machines have the ability of pattern recognition is the direction of scientists' continuous research. The study of pattern recognition is the fundamental way to understand the essence of artificial intelligence. At present, computer vision has been used to identify and count cells. In view of the problems existing in the field of cell counting, such as difficulties in identifying cells, dense cells, viscosity, difficult cell counting, inaccurate judgment and so on, this paper realizes the identification and counting through the technology of computer vision, determines the types of cells, and explores the characteristics of dense and sticky cells, At the same time, based on the characteristics of dense and sticky cells, a color based threshold segmentation algorithm is introduced to optimize the performance of cell classification. At the same time, the recognition and classification of different regions of adhesive cells are carried out, and the adhesive cells in different regions are recognized and judged by the technology of computer vision. Through gray-scale transformation, threshold segmentation, morphological processing and LOG algorithm filtering of medical cell image, the cell region is extracted and counted. The cell counting is realized by calculating the connected domain. The experimental results preliminarily verify the effectiveness of this method, but it still needs to be improved.