Date of Award

Fall 2002

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Electrical Engineering - (Ph.D.)

Department

Electrical and Computer Engineering

First Advisor

Yun Q. Shi

Second Advisor

Nirwan Ansari

Third Advisor

Constantine N. Manikopoulos

Fourth Advisor

Huifang Sun

Fifth Advisor

Anthony Vetro

Abstract

The Internet has become a primary medium for information transmission. The unreliability of channel conditions, limited channel bandwidth and explosive growth of information transmission requests, however, hinder its further development. Hence, research on robust and efficient delivery of video/image content is demanding nowadays.

Three aspects of this task, error burst correction, efficient rate allocation and random error protection are investigated in this dissertation. A novel technique, called successive packing, is proposed for combating multi-dimensional (M-D) bursts of errors. A new concept of basis interleaving array is introduced. By combining different basis arrays, effective M-D interleaving can be realized. It has been shown that this algorithm can be implemented only once and yet optimal for a set of error bursts having different sizes for a given two-dimensional (2-D) array.

To adapt to variable channel conditions, a novel rate allocation technique is proposed for FineGranular Scalability (FGS) coded video, in which real data based rate-distortion modeling is developed, constant quality constraint is adopted and sliding window approach is proposed to adapt to the variable channel conditions. By using the proposed technique, constant quality is realized among frames by solving a set of linear functions. Thus, significant computational simplification is achieved compared with the state-of-the-art techniques. The reduction of the overall distortion is obtained at the same time. To combat the random error during the transmission, an unequal error protection (UEP) method and a robust error-concealment strategy are proposed for scalable coded video bitstreams.

Share

COinS