为了减少分布式视频编码系统解码后重建值与真实值之间误差,提出了基于变换域Wyner-Ziv视频编码最小均方误差(MMSE)重建的一种新的有效重建算法。该算法充分利用视频帧间相关性,对MMSE重建算法积分区间做出调整,当边信处于解码值对应量化区间之内时,在量化区间内利用MMSE重建;当边信息处于解码值对应量化区间之外时,对量化区间做出调整,在改进后的区间利用MMSE重建。实验结果表明,与最佳重建最小均方误差重建算法相比,该算法可以有效提高解码视频的平均PSNR。
In order to reducing the error between the reconstructed value and the ture vale after the distributed video coding system decoding, a new efficient reconstruction algorithm based on transform domain Wyner-Ziv video coding minimun mean square error(MMSE) reconstruction is proposed. It makes full use of the temporal correlation among the video, making an adjustment for the integral interval of MMSE reconstruction, when side information is within the decoded quantization bin, the reconstructed value is derived from using MMSE on the original quantized interval; when the side information is outside the decoded quantization bin, by adjusting the integral interval, the reconstructed value is derived using MMSE on the improved interval. Experimental results show that compared with the optimal the minimum mean square error reconstruction algorithm,the proposed method can improve the average PSNR of the reconstructed video effectively.