Computer surgical simulation has a great deal of potential in medical education and testing. However, there are numerous problems in integrating simulation software technology into a medical curriculum. Review and analysis of the data from the simulation is important to evaluate and assist students. A combination of realistic rendering for good translation of skills to the operating room and illustrative rendering to aid novices can help the simulation system target a wide range of students.
In the context of an ongoing project to develop and improve a temporal bone surgical simulator, this document describes algorithms that address these issues and provides solutions to them. In collaboration with expert surgeons, we have met some of the technological challenges that limit surgical simulation.
Storage and playback of the interactions that users have with the simulation system is achieved via a snapshot technique using forward differences for efficient compression. A technique for realistic rendering of fluid and wet surfaces in a virtual surgical environment using modern graphics hardware is explained. Using a modified distance field technique, we show how to display context around important anatomical structures in segmented datasets. A method of automatic scoring of the users of the simulator is detailed. This method involves partitioning the volume based on proximity to critical structures and then using the Earth Mover's Distance to compare the content of those partitions. Distance fields are also employed for shape analysis techniques to extract features that are used in a visualization system. This system allows expert surgeons to examine and compare the virtual mastoidectomies perfomed by residents during training.