Browse Prior Art Database

EFFICIENCY OF FEATURE DEPENDENT ALGORITHMS FOR THE PARALLEL PROCESSING OF IMAGES

IP.com Disclosure Number: IPCOM000128450D
Original Publication Date: 1983-Mar-01
Included in the Prior Art Database: 2005-Sep-16
Document File: 8 page(s) / 43K

Publishing Venue

Software Patent Institute

Related People

Mudge, T.N.: AUTHOR [+4]

Abstract

In this paper the concept of feature (in)dependent image processing algorithms is defined. A large class of image processing computers characterized by multiple processor-memory subsystems is efficient when dealing with feature independent algorithms but less efficient when dealing with feature dependent algorithms. Typically such machines are required to perform both types of algorithms. This paper is a preliminary attempt to provide a framework within which to model feature dependent algorithms, and to, for example, quantify the inefficiency that can occur when they are executed on the above type of parallel image processors.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 14% of the total text.

Page 1 of 8

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

EFFICIENCY OF FEATURE DEPENDENT ALGORITHMS FOR THE PARALLEL PROCESSING OF IMAGES

T. N. Mudge and T.A. Rahman

CRL-TR-11-83

MARCH 1983

THE UNIVERSITY OF MICHIGAN COMPUTING RESEARCH LABORATORY1 Room 1079, East Engineering Building
Ann Arbor, Michigan 48109
USA
Tel: (313) 763 Efficiency of Feature Dependent Algorithms for the Parallel Processing of Images by
T. N. Mudge and T. A. Rahman

Computing Research Laboratory
Department of Electrical and Computer Engineering University of Michigan
Ann Arbor, MI 48109

Abstract

In this paper the concept of feature (in)dependent image processing algorithms is defined. A large class of image processing computers characterized by multiple processor-memory subsystems is efficient when dealing with feature independent algorithms but less efficient when dealing with feature dependent algorithms. Typically such machines are required to perform both types of algorithms. This paper is a preliminary attempt to provide a framework within which to model feature dependent algorithms, and to, for example, quantify the inefficiency that can occur when they are executed on the above type of parallel image processors.

Keywords -- feature dependent algorithms, image processing, parallel processing.

1. Introduction

The economics of modern digital integrated circuit technology no longer restricts the designers of digital systems to the classical serial interpreter typified by the von Neumann uniprocessor architecture. This trend away from conventional machines is particularly well developed in the field of image processing where the large data sets (64K bytes to 4M bytes per image) and the high processing rates (near term predictions of 1 to 100 billion operations per second have been

1 This work was supported in part by AFOSR Grant No. F49620-82-C-0089. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the funding agencies.

University of Michigan Computing Research Laboratory Page 1 Mar 01, 1983

Page 2 of 8

EFFICIENCY OF FEATURE DEPENDENT ALGORITHMS FOR THE PARALLEL PROCESSING OF IMAGES

made in2) make special purpose machines an economic necessity3 A number of people have proposed/constructed special purpose machines for image processing. These are surveyed in4.

An architectural characteristic of most of these special purpose image processors is a large number of processors working in parallel. Parallel processing is a natural strategy for dealing with the large data sets and high processing rates encountered in image processing applications; furthermore, the nature of the data and the nature of many of the algorithms make parallel processing particularly attractive. The data is usually a large two dimensional array, and many of the low level image processing algorithms can be decomposed into a large number of concurrent neighborhood operations....