Visualization of 3D Scientific Datasets Based on Interactive Clipping

Abstract

We have exploited clipping in combination with multiresolution rendering to visualize large-scale threedimensional (3D) scientific datasets. Our interactive clipping approach involves the dynamic manipulation of a clip plane to expose any cross-section of a given volume data and the subsequent adjustment of the clipped surface to the best view position. The data are rendered as stacks of 2D textures at high-resolution (HR, the original resolution of input data/image) and low-resolution (LR, sampling at reduced-resolution). While clipping operations take place in LR to achieve an interactive frame rate, the best-view position also supports HR. Our approach thus enables a real-time exploration of any interior region of 3D volumetric data at a desired resolution. Its successful demonstration has been carried out for two scientific datasets, which are the simulated electronic density distributions in a crystal and confocal microscopy images of tissues in a plan stem.