Correlation Noise Estimation for Wyner-Ziv Video Transmission over Wireless Network

The rate-distortion (R-D) performance of Wyner-Ziv (WZ) video coding strongly depends on modeling the correlation noise (CN) between the original frame and its side information (SI) frame. The CN model parameters are generally estimated at the decoder in the absence of original video frames. Most of previous methods in the literatures did not take the adverse effects of wireless transmission errors to the estimation of the CN model parameters into account. In this paper, we propose a new scheme of CN model estimation for WZ video transmission over wireless networks with packet erasures and bit errors. The presented scheme can alleviate the influence of transmission errors on the accuracy of CN model estimation by using error concealment for key frames and refinement to the estimated residual frames. Experimental results show that the proposed scheme can improve the R-D performance of WZ video transmission over wireless networks compared with the scheme that have no error concealment for key frames and refinement to the estimated residual frames.