Color Cell Image Segmentation Using Pyramidal Constraint Satisfaction Neural Network

Automatic analysis of medical cell images has been gaining more importance in pharmacology and toxicology research. Segmentation of cell bodies is the crucial step in their analysis since result of segmentation directly affect the following quantitative analysis steps. On the other hand, various smoothing representations of cell bodies are desired to calculate different parameters of the cells. In this work, a new version the Constraint Satisfaction Neural Network (CSNN) algorithm is used to segment the cell images in various details. This is accomplished by incorporating in the CSNN algorithm a multiresolution scheme using pyramid structure.