Date of Award


Publication Type

Doctoral Thesis

Degree Name



Electrical and Computer Engineering


Applied sciences, Compression rate, Rate control, Video coding


Abdel-Raheem, Esam




Advanced video coding standards have become widely deployed coding techniques used in numerous products, such as broadcast, video conference, mobile television and blu-ray disc, etc. New compression techniques are gradually included in video coding standards so that a 50% compression rate reduction is achievable every five years. However, the trend also has brought many problems, such as, dramatically increased computational complexity, co-existing multiple standards and gradually increased development time. To solve the above problems, this thesis intends to investigate efficient algorithms for the latest video coding standard, H.264/AVC. Two aspects of H.264/AVC standard are inspected in this thesis: (1) Speeding up intra4x4 prediction with parallel architecture. (2) Applying an efficient rate control algorithm based on deviation measure to intra frame. Another aim of this thesis is to work on low-complexity algorithms for MPEG-2 to H.264/AVC transcoder. Three main mapping algorithms and a computational complexity reduction algorithm are focused by this thesis: motion vector mapping, block mapping, field-frame mapping and efficient modes ranking algorithms. Finally, a new video coding framework methodology to reduce development time is examined. This thesis explores the implementation of MPEG-4 simple profile with the RVC framework. A key technique of automatically generating variable length decoder table is solved in this thesis. Moreover, another important video coding standard, DV/DVCPRO, is further modeled by RVC framework. Consequently, besides the available MPEG-4 simple profile and China audio/video standard, a new member is therefore added into the RVC framework family. A part of the research work presented in this thesis is targeted algorithms and implementation of video coding standards. In the wide topic, three main problems are investigated. The results show that the methodologies presented in this thesis are efficient and encouraged