Approaches to the construction of cognitive linguistic–image models of knowledge representation for medical intelligent systems

Abstract

Traditional linguistic–logical models in intelligent medical systems do not provide an exact match between the cognitive representation of an expert and the linguistic representation in a knowledge base, with regard to holistic visual images of diseases that meet certain definitions. The appearance of a patient is often important in medical diagnostics, in particular in hereditary diseases. However, we should note that there are no elaborated methods and tools to support special models that make it possible to use knowledge presented in verbal and figurative forms within the same formalism. Current national and foreign research and development in this field are mostly aimed at solving specific problems focused on a particular side of the problem. Clinical genetics is an area where the creation of systems based on a combination of verbalized and figurative knowledge can provide a qualitatively new level of diagnostics. Awareness of these factors results in the fact that the developed countries have become highly interested in medical-image analysis in recent years. However, these developments are not focused on the use of expert knowledge. One possible solution could be to construct a hybrid linguistic–image framework or ontological models that integrate qualitative and visual–image indicators for the formal description of diagnostically important clinical manifestations of diseases.