Development Of Cell Counter With Auto Focus Based On PCNN

In this paper,a kind of cell counter based on Pulse Coupled Neural Network(PCNN)which can focus automatically is designed.The cell counter is used to count the suspended living cells in the bright field.The cell counter system consists of optical system,software system and electromechanical control system.In the optical system,the Kohler illumination is used for illumination,and the relevant parameters are calculated;the PW method is used for the initial structure design of the microscope objective,and then the initial structure parameters are optimized and simulated by ZEMAX software.Cell counting algorithm is the core part of this paper.Due to the characteristics of ignition capture of PCNN neurons model,we use PCNN model to segment the central highlight area,which can be treat as the mark of living cells,of suspended cells.And then get the number of connected components from the cell mark image by the binary image connected component acquisition algorithm,it is the very number of the cells.Experiments show that the algorithm has high precision and recall rate for the cell image without stratification,but for the cell image with stratification,the cells in the lower layer of the solution cannot be captured by PCNN model,and these cells will be lost.Through the comparison of several commonly used image sharpness evaluation functions,it shows that these kinds of sharpness evaluation functions can not directly search the best focus position by the cell image.But in the experiment,it is found that it is feasible to search out the best focusing position of cell counter by NRSS sharpness evaluation function in standard PS small ball sequence image.