Knowledge-based interpretation of cranial MR images

A computerized system is described that can be used to evaluate an MR tomogram automatically to support clinical identification of anatomic and pathologic structures and to aid in planning MR measurements. Knowledge from three domains is used for the interpretation of an MR image: nosologic knowledge, knowledge of MR imaging parameters, and anatomic and morphologic knowledge. Nosologic information is used to generate hypotheses about possible pathologies and their locations, based on the signs and symptoms of the patient. With this information, a sequence of interpretation modules, each able to detect substructures in already detected structures with the aid of techniques from image processing, pattern recognition, and artificial intelligence, is generated and executed