Distributed Video Coding with Dynamic Virtual Channel Model Estimation

Distributed video coding (DVC) is a new video coding paradigm, based on Slepian&Wolf´s and Wyner&Ziv´s information theories. A practical case of DVC is Wyner-Ziv codec based on turbo code. One of the most significant works in Wyner-Ziv codec is the estimation of virtual channel model between the side information and the original frame. Better estimation leads to better coding efficiency of DVC. In this paper, the dynamic property of the virtual channel model is discussed in different levels of granularity. A new dynamic virtual channel model estimation algorithm in decoder is proposed. Experimental results show that the proposed algorithm leads to significant improvement in the rate-distortion performance of DVC.