This research involves the study and development of an integrated framework of pre- and post-filtering along the data block boundaries in block-based image/video communication systems. The novel framework can be thought of as a generalization of and provides a significant improvement for the traditional block DCT coding approach in JPEG and MPEG. The investigation is particularly geared toward low-complexity image/video coding and communication algorithms or systems in resource-constrained, speed-critical, and real-time applications for wireless hand-held devices.
The research is expected to have a significant impact on practical signal processing, coding and communication systems. Coding efficiency is greatly improved by taking into account inter-block spatial correlation and by adapting operators dynamically based on local signal contents. At the decoder, perceptual reconstruction quality is enhanced. Blocking artifacts are mitigated with aggressive post-filtering along the block boundaries whereas ringing artifacts are greatly reduced by adaptively varying the operator size. Robustness and data recovery capability in noisy channel environments are also improved by spreading information spatially and by dynamically switching interpolative post-filters based on the communicating environment and the neighboring data received. Finally, on the practical aspect, small local operators lend themselves nicely to parallel computing, and fast, VLSI-friendly, possibly multiplierless, implementations. Existing block-based infrastructure can be kept intact since pre-/post-processing operates outside in the spatial domain. The investigator explores pre-/post-processing operators that can maintain compatibility with the huge collection of JPEG and MPEG bit-streams currently available on the web and in archiving digital libraries.
Level of Effort Statement:
At the recommended level of support, the PI will make every attempt to meet the original scope and level of effort of the project.
