Wavelets and Recent Developments in MPEG Video

Video coding research and standardization efforts have traditionally focused on techniques and architectures that provide the maximum coding efficiency with the use of the latest advances in signal approximation and compression theory.

In a shift from the conventional paradigm, the MPEG standardization committee recently decided to investigate wavelet-based video coding technologies that, although promising to deliver small advances in terms of compression efficiency over the state-of-the-art, provide a compressed bitstream with a rich set of features for efficient manipulation and protection of the encoded information. This efficient manipulation centralizes around the capability for optimized repetitive adaptation of the compressed video data to different decoder resolutions, frame-rates and picture-quality levels from one single compressed bitstream. This transcoding-free multiple-adaptation mechanism is traditionally termed as full-scalability.

In this talk the architectures and signal processing tools that lead to fully-scalable video-compression methods will be presented. The essential research advances that paved the path for this new phenomenon of video-coding systems with multiple adaptation in time/space/quality-sense and cutting-edge performance will also be explained. Relevant results are presented to experimentally substantiate each case. Many of the remaining open problems are identified and predictions are made over the future technological developments until this research reaches mainstream commercial applications.

Hide picture