Wyner-Ziv视频编码中边信息估计算法改进*

与传统视频编码方法相比，DVC（distributed video coding）在编码性能方面还存在着较大差距。边信息估计是其中的关键技术之一，在很大程度上决定着编码效率。为缩短性能差距，改善边信息估计效率，提出一种针对像素域Wyner-Ziv视频编码系统的改进算法，在解码端改善了关键帧之间的运动矢量获取以及采用重叠块运动补偿来生成边信息。通过对大量测试序列的实验，验证了改进算法的率失真性能得到改善。

Improving on side information estimation for Wyner-Ziv video coding

DVC had the substantial performance gap to well known state-of-the-art coders. Side information estimation was one of the key technologies for DVC, which decided the code efficiency heavily. In order to shorten the performance gap an improve the efficiency of side information estimation, provide the proposed scheme for pixel-domain Wyner-Ziv coding, which modified the way of acquiring motion vector among successive key frames on decoder. Based on the achieved motion vector, generated the side information with the overlapped block motion compensation. Extensive experimental results show that the improvement on rate-distortion (RD) is achieved with the proposed method.