Study Of The Key Technologies For Distributed Video Coding

With the development of wireless network, more and more mobile video terminals are used in multimedia communications. Since these wireless devices are always limited in computation capability and energy resources, the encoder should be low-complexity and error resilient with high compression efficiency. The traditional video compression standards such as MPEG and H.26x could not used in the wireless network with high error probability and limited resources. To tackle the problem, the scholars home and abroad proposed the novel video compression method: distributed video coding technology.Slepian-Wolf and Wyner-Ziv theorem are basic principles of distributed video coding. Slepian-Wolf theorem is for lossless coding and Wyner-Ziv theorem is the loss extension of Slepian-Wolf theorem. First, the thesis presented the basic principles, and the two typical distributed video coding systems, proposed by Stanford University and Berkeley, California University, are introduced in detail. Then a comprehensive analysis on the characteristics of distributed video coding is given such as low-complexity encoder and error resilience.The thesis then presented a research on the key technologies such as rate control and side information generation. On the one hand, a low-complexity rate control scheme is proposed at the encoder to avoid the feedback channel. On the other hand, side information generation method based on motion compensation and interpolation using the forward and backward neighboring frames is proposed to precisely estimate side information. A practical distributed video coding system in transform domain based on the above two algorithm is finally constructed. The experimental results show that compared with H.263 intra coding, the coding efficiency is significantly improved and the reconstruction quality is obviously enhanced. Furthermore, the encoding speed is considerably improved compared with H.263 inter coding. The proposed distributed video coding system can be applied to the video transmission systems for low-complexity encoding terminals.