Improvements in wavelet-based rate scalable video compression
Abstract
Delivery of video in the presence of bandwidth constraints is one of the most important video processing problems. Most current compression techniques require that parameters, such as data rate, be set at the time of encoding. However, it is difficult to predict the traffic on a network when video is to be delivered. Compression techniques that allow the change of the compression parameters at the time of decoding are very valuable due to the flexibility they provide. These techniques are said to be “scalable.” Rate scalability, the capability of decoding a compressed image or video sequence at different data rates, is the one of the most important modes for video streaming over packet networks such as the Internet. SAMCoW is a rate scalable, wavelet-based video compression algorithm developed at Purdue University in 1997 by Shen and Delp. SAMCoW is fully and continuously rate scalable, and has performance similar to MPEG-1, MPEG-2, and H.263 at comparable data rates. In this thesis, we investigate several extensions to SAMCoW. In particular, we investigate the use of advanced video coding techniques, preprocessing, postprocessing, and rate-distortion theory, in order to develop a framework for efficiently encoding intracoded (I) and predictive error (PE) frames as part of SAMCoW. These extensions are known as SAMCoW+ . We also investigate the use of digital signal processors for real-time video processing. We demonstrate the use of a Texas Instruments TMS320C6201 for real-time error concealment of digital video streams.
Degree
Ph.D.
Advisors
Salama, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.