Computational analysis of real-time MPEG-2 video compression and its hardware implementation

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Electrical Engineering and Computer Science


C. Y. Roger Chen


motion estimation, Electrical engineering, Computer science

Subject Categories

Computer Engineering


The establishment of video compression standards and the extensive application of compression technology necessitate the consideration of a practical solution for real-time MPEG encoding. Due to the high computational requirements of the MPEG encoding algorithm, hardware-based solutions are at the present time the only viable alternatives.

A detailed complexity analysis of the MPEG-2 main profile and main level encoding algorithm is performed in this work. Based upon this analysis, the investigation of the internal structure of the state-of-the-art multimedia video processor (MVP) and the mapping of the encoding functions on the MVP, an architecture for real-time encoding/decoding of the single constraint MPEG-2 syntax is developed. A major feature of this architecture is its programmability, a feature which may be utilized for other computationally extensive algorithms.

The encoding problem is a constrained optimization problem. The results of statistical studies performed towards this end and the development of a new motion estimation algorithm are employed in developing a more feasible architecture for the MPEG-2 main profile and main level real-time encoding.


Surface provides description only. Full text is available to ProQuest subscribers. Ask your Librarian for assistance.