CAREER: Error Correction and Concealment for Wireless Video: An Integrated and Adaptive Approach

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

0133221 Altunbasak, Yucel GA Tech Res Corp As wireless networks become more pervasive, there is a great need for wireless multimedia with quality and reliability levels that are characteristic of wired networks. It is clear that wireless communications will include image, video and data applications. Furthermore, spectral limitations dictate that transmission resources be used efficiently. To this end, image, video and other mixed media encoding algorithms need to be designed to overcome the types of channel errors that are likely to occur. Noting the wireless channel characteristics and the requirements of real-time video applications, the investigator offers solutions that are novel in two aspects. First, the investigated methods are adaptive with respect to variations in bandwidth and channel error statistics. This flexibility should prove to be critical in heterogeneous wireless environments. Second, the error correction and concealment methods are not only technology advances by themselves, but they also function together in innovative ways. More specifically, the contributions of this research include: (1) a coordinated multiple description coding and error concealment procedure where the smoothness of the video signal helps to compensate for the loss of descriptions, (2) a multiple description coder that adapts to variations in data rate and packet loss statistics, (3) a new multi-frame boundary matching algorithm that utilizes the boundary smoothness property not only in the decoded frame but also in the successive frames, thereby minimizing the error propagation, (4) a new multi-frame blocking artifact reduction algorithm that uses the spatial correlations that exist between the successive frames to define constraint sets at multiple frames and provides a projections onto convex sets based solution, (5) a transport protocol that is specifically designed for real-time multimedia communications, and (6) a testbed to integrate all the aforementioned methods into a system-level wireless video streaming solution. On the educational front, the investigator plans to develop a comprehensive program in multimedia signal processing and communications. The investigator will make every attempt to meet the original scope and level of effort of the project.