Multi-segment and multi-stage projected tetrahedra

Projected tetrahedra is a popular volume rendering technique to find the interior features of tetrahedral datasets. It is a view-dependent technique and needs to sort the tetrahedra before rendering. Sorting is time-consuming and affects the efficiency of the visualization heavily. To solve this problem, our approach partitions datasets into multiple segments in space and divides the visualizing procedure into multiple stages in time. Multiple segments help sorting in parallel without write access violations while multiple stages can satisfy different requirements and increase users' cognition. Cloud computing can help this process because datasets are large and sorting is complex. The experiment results show that our visualization approach for tetrahedral datasets is an exact and efficient approach, which accords with cognitive laws.