Browse Prior Art Database

Integrated Computer Architectures for Image Processing Database Management

IP.com Disclosure Number: IPCOM000131578D
Original Publication Date: 1983-Jan-01
Included in the Prior Art Database: 2005-Nov-11
Document File: 9 page(s) / 36K

Publishing Venue

Software Patent Institute

Related People

Kai Hwang: AUTHOR [+4]

Abstract

With VLSI technology, we can integrate image processing, pattern recognition, and database management to produce a cost-effective computer system for advanced automation and machine intelligence.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 13% of the total text.

Page 1 of 9

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

This record contains textual material that is copyright ©; 1983 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Contact the IEEE Computer Society http://www.computer.org/ (714-821-8380) for copies of the complete work that was the source of this textual material and for all use beyond that as a record from the SPI Database.

Integrated Computer Architectures for Image Processing Database Management

Kai Hwang and King-sun Fu,

Purdue University

With VLSI technology, we can integrate image processing, pattern recognition, and database management to produce a cost-effective computer system for advanced automation and machine intelligence.

L image analysis, the use of digital computers for pattern recognition and image processing, on- line im agery data need to be stored on disks and quickly re trieved. For a pictorial information system to be effec" five, then, it must efficiently manage and analyze im" agery data. Special- purpose computer architectures for processing pictorial information should integrate both pattern analysis and image database management capabilities into a unified design aimed at the development of a real-time, interactive computer system for multidimensional information processing. This integrated approach is already being implemented in the Pumps project at Purdue University.'~3

A typical image analysis system consists of four processing stages as depicted in Figure I. The preprocessing stage includes image operations like smoothing, enhancement, restoration, edge detection, and segmentation. Raw images are reduced to segmented patterns in this initial stage. The second stage is for feature extraction, which further reduces the segmented image to a small set of feature vectors. Clustering techniques can be applied at this stage. The third stage is for pattern classification, which recognizes the membership of extracted features among known pattern classes. The fourth stage is for structural analysis and interpretation, to produce a concise description and interpretation of pattern information.4~7

Conventional single instruction, single data stream computers are designed primarily to process one-dimensional strings of alphanumerical data. To process multidimensional information on SISD computers requires image coding and picture transformation, such as projection and registration. Sequential machines cannot efficiently exploit the parallelism inherent in most pattern

recognition and image processing, or PRIP, operations. On the other hand, large parallel computers, such as single instruction, multidata stream array processors and multi- instruction, multidata stream multiprocessors, are not necessarily cost-effective in implementing simple and repetitive image operations over very large and, sometimes, dynamically changing image databases.

(Image Omitted: Figure 1. Processing stages of an image analysis computer sy...