Mathematical Linguistics in Cognitive Medical Image Interpretation Systems

The task of carrying out semantic searches for useful information was solved to some degree for textual data. Unfortunately, medical image analysis has demonstrated that the issue of searching for useful semantic information, on the basis of content, is an insolvable task, in practice. We have acquired a certain expertise in using automatic techniques for retrieving features that define the semantic content of visual data. Everything seems to suggest that the techniques of syntactic pattern analysis which are useful in computer-aided medical diagnosis can also be very helpful in tasks involving automatic understanding procedures used in medical image interpretation systems. Therefore, the topic of this paper will be the presentation of application possibilities for graph EDT grammars as well as context-free grammars as cognitive techniques supporting medical pattern semantic interpretation. Those mathematical linguistic mechanisms will be used for the automatic generation of syntactic descriptions of the contents of the analyzed types of medical images originating from pancreas ERCP examinations, spinal cord visualization and wrist radiograms. The semantics of those patterns define the type of incorrectness of the organ shown on it. Attention will be focused on the methods of cognitive analysis with respect to medical data. So, the languages of shape description and proposed grammars mainly allow for the creation of syntactic descriptions of selected anatomic organs together with a definition of the semantic meaning of the changes in their shapes. Such descriptions will allow us to create a semantically-oriented representation of visual data describing pattern features which are useful in medical decisions and computer-aided diagnostic systems.