Publication Date

1990

Document Type

Dissertation/Thesis

First Advisor

Miller, Gerald D. (Professor of electrical engineering)

Degree Name

M.S. (Master of Science)

Legacy Department

Department of Electrical Engineering

LCSH

Image processing--Digital techniques; Data compression (Computer science)

Abstract

This thesis extrapolates and expands upon an existing image data compression algorithm called the “Predictive Ordering and Linear Approximation.” The attractiveness of this algorithm is it’s reduced computational demands relative to other image compression algorithms. This algorithm supports compression ratios in the order of 8:1 while reducing the image reconstruction’s absolute mean error to approximately 5%. The work presented here includes an analysis of an improvement to the algorithm as well as the presentation of several sorting methods and hardware structures for real time execution. The ramifications of replacing the linear approximation of the compression algorithm with a quadratic approximation are explored. Specific images have exhibited absolute mean error reductions of 15 to 20% while increases in the number of bits required to represent a picture element of 5% are supported. Because of the algorithm’s real time execution dependency upon sorting efficiency, the classical selection, exchange and merge sort techniques were implemented as software modules and executed on two simulated hardware architectures. These architectures are DSP based and are included for real time computational analysis purposes. Results indicate that a speedup factor of up to 5 .0 can be obtained during compression/ regeneration of the image. The approach presented in this thesis looks very promising for a large set of those real time digital image transmission applications which heretofore have been unachievable.

Comments

Includes bibliographical references (page 47)

Extent

94 pages

Language

eng

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Share

COinS