The goal - so far not entirely realized - of synthetic environments is to provide a realistic experience for their users. In particular, these environments can be used for active collaboration on engineering designs and medical applications. This project attacks some of the most important outstanding problems in the field of real-time synthetic environments:
1) Develop a scalable system to achieve real-time walkthrough of very large CAD models with high-accuracy rendering of spatial arrangements.
2) Develop parallel rendering algorithms that take into account emerging graphics architectures with multiple graphics pipelines and general-purpose processors.
3) Develop incremental algorithms to allow interactive updates to large dynamic environments, allowing objects to be moved, added, and removed in realistic manners.
Among the algorithms that the project will explore are visibility culling, multi-resolution methods, model simplification, image-based rendering,3-D image warping, multi-pass rendering, and efficient updates for model hierarchies, occlusion databases, and level-of-detail generation. These advances will be tested on real-world models, including a 13 million-element polygonal model of an electric generating station.
