Share Email Print
cover

Proceedings Paper

Parallel architecture for multidimensional image processing
Author(s): Raj S. Acharya; Jean Hecquard
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

In order to deliver the computational power required by real-time manipulation and display of multidimensional objects, we present a massively parallel octree architecture, based on a new Interconnection Network, the Partitionable Spanning Multibus Hypercube (PSMH). Its goal is, to use one Processing Element per obel (object element), as opposed to one Processing Element per voxel (volume element). The underlying idea of the PSMH, is to take advantage of the data hierarchical ordering to reduce the computational cost. As a basic tool, we first derive a routing algorithm for the case of an object shift. Its running time is of order O(max(n3, m)), for an 8n PSMH, where m is the message length in bits. As we do not consider voxels but obels, we design a compaction algorithm, which meets the routing requirements. We get a compression ratio of O(2n). This is followed by a parallel neighbor finding technique, to account for the compaction in the routing operations.

Paper Details

Date Published: 1 July 1990
PDF: 23 pages
Proc. SPIE 1246, Parallel Architectures for Image Processing, (1 July 1990); doi: 10.1117/12.19574
Show Author Affiliations
Raj S. Acharya, SUNY/Buffalo (United States)
Jean Hecquard, SUNY/Buffalo (United States)


Published in SPIE Proceedings Vol. 1246:
Parallel Architectures for Image Processing
Joydeep Ghosh; Colin G. Harrison, Editor(s)

© SPIE. Terms of Use
Back to Top