Volumetric Depth Peeling on Graphics Hardware
Cory Quammen · 2006
Figure 1: Single view of an MRI scan of an elbow joint both without volumetric depth peeling (left) and with volumetric depth peeling (right). Uninteresting data from skin and muscle occludes the elbow joint on the left. Volumetric depth peeling culls the uninteresting data in a view-dependent way, revealing the elbow joint. The increasing programmability of modern graphics hardware enables acceleration of rendering methods other than traditional rasterization. In this paper, we discuss the implementation of a ray casting-based volume renderer on graphics hardware using vertex and fragment programs. We also show how ray casting on hardware enables easy implementation of volume rendering extensions that make use of per-ray state, using volumetric depth peeling as an example. We also compare the performance of the hardware and software implementations of volumetric depth peeling.