| Miriah Meyer | |
Hanspeter Pfister | |
Charles Hansen | |
Chris Johnson |
| SCI Institute University of Utah |
|
MERL Cambridge, MA |
|
SCI Institute University of Utah |
|
SCI Institute University of Utah |
Abstract:
While low cost PC graphics hardware has proven valuable for volume rendering, large datasets continue to overwhelm the capabilities of the graphics cards, reducing the interactivity of volume rendering utilizing such hardware. We present a novel, image-based approach to volume rendering that can render arbitrarily large datasets interactively on current graphics hardware. Our method is independent of the volume rendering system and the dataset representation, and allows for exploration of the interior structure of the volume. The process consists of three main steps, each of which can be run independently. In the first step, a set of ray slices of the data are produced from multiple viewpoints using a volume render, and a geometric proxy surface bounding the volume is defined. Next, the ray slices and geometric proxy are processed to compute a set of key views. Finally, the key views and proxy surface are rendered interactively as opacity light fields on current graphics hardware. The user can change the proxy surface to reveal the interior structure of the volume data. Our method has been tested on a variety of volume datasets and these results are presented in this paper.