Flexible and Robust Coding of Images and Video

University Of Washington, Seattle WA

Investigators

Abstract

Proposal # 0104800 PI: Eve Riskin U of Washington In the not-too-distant future, downloading images and video to a handheld device will be as commonplace as viewing an image on a Web page is today. To enable this, high quality, flexible, and robust image and video compression algorithms will be required. Recently, Group Testing for Wavelets (GTW), a new type of wavelet coder based on group testing was developed. It offers competitive performance to the best compression coders available today and with further development, GTW could outperform them significantly. This research involves developing new algorithms for robust and flexible coding of images and video. Whereas GTW is used as a motivating example, the research is applicable to many compression algorithms and in many scenarios. First, a set of problems related to GTW will be explored. This includes applying GTW to wavelet packet decompositions and reduced complexity transforms; extending it to video and the new area of progressive geometry compression; and developing a theory for its strong performance. Next, methods for applying forward error correction to compressed data in a progressive manner will be investigated. Finally, unequal frame expansions will be used to recover from packet loss. The research will have practical use in many arenas including the Internet and wireless communications and will be included in a new undergraduate course on data compression, to be taught jointly by the two principal investigators.