An New Automatic Nucleated Cell Counting Method With Improved Cellular Neural Networks (ICNN)

The number of nucleated cells in a bone marrow slice image is an important sign of the degree of hyperplasia. But because of the complex components in the bone marrow slice image, there are only a few methods available for cell detection or segmentation with poor effect. This paper focuses on this issue. A new automatic nucleated cell counting method based on improve cellular neural networks (ICNN) is proposed. We improve the CNN output function and practically design all the templates, then prove the stability. The process details are also presented. Experimental results show good performance. ICNN can detect almost all nucleated cells. Its running speed is expected to be comparatively high due to the easy hardware implementation and high speed of CNN