Exploiting 3d part-based analysis, description and indexing to support medical applications

Multi-modality is crucial to handle knowledge, medical data and patient-specific information in an integrated fashion: in the course of their work, indeed, clinicians analyze a large amount of information about interrelated layers such as anatomy, kinematics, dynamics, mechanics and physiology. Much of the information related to these levels is intrinsically 3D and we believe that the adoption of 3D part-based annotation and content-based indexing will open up new ways to integrate and interact with medical information. In this paper, we will focus the attention on content-based analysis of 3D medical data and discuss related issues and trends, based on two software tools: the ShapeAnnotator and RheumaSCORE. In the illustrative scenario of the Rheumatoid Arthritis we will provide hints for even more informative Computer Aided Diagnosis systems for clinical support.