This paper deals with theoretical and application-related issues connected with the possibilities of creating intelligent bio-medical and multimedia systems based on graph linguistic formalisms, originating from AI methods. Such systems can be directed at tasks supporting medical diagnosis, multimedia indexation as well as classification and cognitive semantic perception of the content of medical images. The said possibilities can be obtained owing to the use of structural and semantic reasoning processes in such systems, conducted pursuant to linguistic rules of appropriately defined graph and sequential grammars, describing the semantics of selected MRI and RTG images.