The use of new staining methods and the large amount of investigations for human chromosomes has created a need for a useful automated chromosome analysis. Because of the lack of accuracy of the existing systems, a new approach is given. The problems in the different stages of the classification process are described. A comparison of the decision-statistical, syntactical, and interpretative approach is discussed. A system realization for the total karyotyping of chromosomes, based on the interpretative approach, is described in more detail. Guided by one of the models in the world model structure, a current description is recognized in a three-phase interpreter.