Compression-Based Ray Casting of Very Large Volume Data in Distributed Environments

This paper proposes a new parallel/distributed raycasting scheme for very large volume data that can be effectively used in distributed environments. Our method, based on data compression, attempts to enhance the rendering speedups by quickly reconstructing voxel data from local memory rather than expensively fetching them from remote memory spaces. Our compression-based volume rendering scheme minimizes communications between processing elements during rendering computation, hence is very appropriate for both distributed-memory multiprocessors and PC/workstation clusters, where the relatively high communication costs often hinder efficient parallel/distributed processing. We report experimental results on both a Cray T3E and a PC/workstation cluster for the Visible Man dataset.