Research On The Theory And Algorithm Of Distributed Video Coding

Distributed source coding is a new paradigm for source compression, based on Slepian-Wolf's and Wyner-Ziv's information-theoretic results from the 1970s. Although the history of Slepian-Wolf and Wyner-Ziv theorem is more than 30 years, the practical algorithm for distributed source coding is proposed in this century.Leveraging the features of distributed source coding, researchers do a lot of research on distributed video coding, and design a series of distributed video coding systems that represent a fundamental paradigm shift from the conventional video coding methods. The conventional video coding comprises of a computationally intensive encoder and a lightweight decoder. But distributed video coding shifts the computational burden from the encoder to the decoder without compromising the theoretical compression performance and offers the feature of naturally robustness against transmission errors. Currently, the rate-distortion performance of pratical distributed video coding system is superior to conventional intraframe coding, but there is still a gap relative to conventional motion-compensated interframe coding.Distributed video coding is well-matched to applications involving wireless multimedia transmission (such as 802.11 or cellular or video sensor networks) which are characterized by the requirements of (i) computational power efficiency at mobile terminals due to battery life considerations; (ii) high compression efficiency due to scarcity of wireless bandwidth; and (iii) robustness to channel loss due to wireless medium.This thesis presents the fundamental concept and theory of distributed source coding, and gives out the relationship of distributed source coding and channel coding. The author gives the fundamental implementation and the state-of-art research results of distributed video coding, and conclude the comparition of distributed video coding and prediction based video coding from the information theoretical aspest. Based on this work, the author designs a distributed video coding system and look into the design principal and performance of each module of this system.