Interpretation of Medical Images Based on Ontological Models

It is presented a concept of using ontological models as a form of medical knowledge representation for computer-aided interpretation of medical images. The models are based on hyper-relations linking the concepts of visual objects, visualized medical objects and related non-visualized real objects. It is shown that due to the algebra of hyper-relations such ontological models provide a possibility to describe large variety of situations being of particular interest in image interpretation tasks. It is defined an association area of a given concept. In the case of visual objects this makes possible characterization of a class of questions that within the given ontological model can be posed and answered by interpretation of the visual objects under examination.