Accelerating Raycasting Utilizing Volume Segmentation of Industrial CT Data

Abstract

We propose a flexible acceleration technique for raycasting targeted at industrial CT data and the context of material deficiency checking. Utilizing volume segmentation that is typically employed for object analysis, GPU raycasting can be accelerated significantly using a novel data structure that is integrated into the volume to improve the responsiveness for the interactive, visual inspection of high-resolution, high-precision data. Our acceleration approach is designed to cause no extra texture lookups and to produce only marginal computational and storage overhead. Despite the fact that the data structure is integrated into the volume, the graphics card's hardware can still be used for trilinear interpolation of density values without producing incorrect results. The presented method can further easily be utilized in combination with out-of-core approaches and distributed volume rendering schemes.