InVIS - Interactive Visualization of Medical Data Sets

Abstract

Interactive rendering of large data sets requires efficient algorithms and accelerated graphics hardware. Both efficiency and speed can be improved, but none of these alone guarantees interactive response time. If a scene is too complex, performance decreases, and neither faster algorithms nor speeding up hardware can ensure interactive performance. This means that certain timing characteristics need to be incorporated in order to obtain interactive behavior.In this paper we propose a new approach to interactive rendering by introducing a rendering pipeline with special timing predicates. We use this technique for applications in the field of medical imaging. Large data sets derived from CT or MRI scans and CAD designs must be rendered in real-time. The system should provide immediate feedback to the user. Interactive behavior enables the user to manipulate and adjust the image straight on demand.