“Chromosome”: A rule-based image analysis system for chromosome classification

A fully automatic system is presented for the analysis and the classification of human chromosomes. Its architecture is based on two cooperating sub-systems: a rule-based system including two types of knowledge (cytogenetics, image analysis) and a tool box containing the low level processings such as pattern recognition and classification methods. The multi-agents design of this system makes further developments easier (modularity) and allows to generalize it to other image analysis applications.