: In multimodal image-guided surgery, surgeons have to manage large amounts of information (i.e. images) to plan and perform the surgical procedure. To facilitate management we attempted to model the procedure as a set of distinct steps with a UML class diagram. To each step we assigned a list of relevant image entities extracted from multimodal images and an action describing the purpose of the step. Each image entity was attributed a role that may differ with the step. Our generic information model helps surgeons anticipate surgery, and may contribute to develop simulation and even teaching tools. It may also be used to adapt visualization and interaction modes in neuronavigation. Finally it provides a framework to model the knowledge involved in neurosurgical procedures in the context of image-guided surgery of the central nervous system.
Pierre Jannin, M. Raimbault, Xavier Morandi, E. Se