Examination of Chaotic Attributes in MPEG Video Sources

by

Denny Lee

M. Eng.,  1995

Abstract

As Asynchronous Transfer Mode (ATM) continues to gain momentum in offering an unify networking technology, the realization of multimedia communications is finally taking shape. Central to this revolution in information exchange is the ability to transport real-time video data. This report provides a study on the examination of chaotic behavior of video sources compressed by the MPEG algorithm. Time Series models under studies are generated by commercially available movies with different post-compression characteristics. Basic statistical analysis on these models are done by using CDF, PDF, FFT and auto acceleration. Furthermore, an attempt to visualize the attractor of the varies time series is presented. The report continues to analyze the chaotic attributes of these MPEG video sources by using the method of Dimension and Lyapunov Exponent. Varies different source models and classical chaotic models are used throughout the report to compare against the experimentally collected that the experimentally collected MPEG time series are not chaotic in nature.

Back to the list of the graduate students