GRAFIP: A FRAMEWORK FOR THE REPRESENTATION OF HEALTHY AND PATHOLOGICAL CEREBRAL INFORMATION

This paper presents a contribution to the large problematic of integrating medical image-based information into a structured framework (such as electronic patient records or anatomo-functional databases). In neuroscience, the complexity of the cerebral anatomy, the wealth of information embedded in imaging data, as well as the difficulty of their interpretation, can benefit from the use of a structural brain model representing prior generic knowledge, which includes information on anatomical structures and their spatial relations. In this paper we describe a novel generic brain model, based on graph representations, and an instantiation procedure for individual patients, based on image segmentation. A complete patient-specific modeling framework is proposed that can be integrated into powerful computational tools to assist image data reviewing, diagnosis and therapeutic patient follow up