Smart hardware-accelerated volume rendering

摘要: For volume rendering of regular grids the display view-plane aligned slices has proven to yield both good quality and performance. In this paper we demonstrate how merge most important extensions original 3D slicing approach, namely pre-integration technique, volumetric clipping, advanced lighting. Our approach allows suppression clipping artifacts achieves high while offering fiexibility explore data sets interactively with arbitrary clip objects. We also outline utilize proposed for segmented sets. Moreover, increase by implementing efficient over-sampling pixel shader consumer graphics accelerators. give prove that at least 4-times is needed reconstruct ray integral sufficient accuracy even pre-integration. As an alternative brute-force propose a hardware-accelerated caster which able perform only where gain additional speed early termination space leaping.