Award Date

1-1-1999

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Committee Member

Evangelos Yfantis

Number of Pages

157

Abstract

The reduction of data necessary for storage or transmission is a desirable goal in the digital video and audio domain. Compression schemes strive to reduce the amount of storage space or bandwidth necessary to keep or move the data. Data reduction can be accomplished so that visually or audibly unnecessary data is removed or recoded thus aiding the compression phase of the data processing. The characterization and identification of data that can be successfully removed or reduced is the purpose of this work. New philosophy, theory and methods for data processing are presented towards the goal of data reduction. The philosophy and theory developed in this work establish a foundation for high speed data reduction suitable for multi-media applications. The developed methods encompass motion detection and edge detection as features of the systems. The philosophy of energy flow analysis in video processing enables the consideration of noise in digital video data. Research into noise versus motion leads to an efficient and successful method of identifying motion in a sequence. The research of the underlying statistical properties of vector quantization provides an insight into the performance characteristics of vector quantization and leads to successful improvements in application. The underlying statistical properties of the vector quantization process are analyzed and three theorems are developed and proved. The theorems establish the statistical distributions and probability densities of various metrics of the vector quantization process. From these properties, an intelligent and efficient algorithm design is developed and tested. The performance improvements in both time and quality are established through algorithm analysis and empirical testing. The empirical results are presented.

Keywords

Audio; Audio Compression; Data Reduction; Dimensions; Image; Motion Detection; Reduction; Space; Vector Quantization; Video Compression

Controlled Subject

Computer science; Electrical engineering

File Format

pdf

File Size

4444.16 KB

Degree Grantor

University of Nevada, Las Vegas

Language

English

Permissions

If you are the rightful copyright holder of this dissertation or thesis and wish to have the full text removed from Digital Scholarship@UNLV, please submit a request to digitalscholarship@unlv.edu and include clear identification of the work, preferably with URL.

Rights

IN COPYRIGHT. For more information about this rights statement, please visit http://rightsstatements.org/vocab/InC/1.0/


COinS