分布式视频编码中拉普拉斯-柯西混合分布相关噪声-电子与信息学报.PDFVIP

第 34 卷第 10 期 电 子 与 信 息 学 报 Vol.34No.10 2012 年 10 月 Journal of Electronics Information Technology Oct. 2012 分布式视频编码中拉普拉斯-柯西混合分布相关噪声模型研究 杨春玲 谢小兰* (华南理工大学电子与信息学院 广州 510640) 摘 要：该文通过研究变换域分布式视频编码中原始 Wyner-Ziv(WZ)帧与相应边信息的残差系数特性，发现大残 差和小残差系数统计分布与传统的拉普拉斯分布存在一定偏差。为了减少这种差异，提出一种拉普拉斯-柯西混合 分布(LCMD)相关噪声模型及其参数估计算法。该混合模型利用改进的拉普拉斯分布描述小残差系数的分布，采用 柯西分布描述大残差系数。实验结果表明该文提出的混合模型能较精确地描述 WZ 帧和边信息间的残差系数分布， 从而有效地改善了变换域分布式视频编码的率失真性能，并减少系统解码端计算复杂度。 关键词：分布式视频编码；拉普拉斯-柯西混合模型；拉普拉斯分布；柯西分布 中图分类号：TN919.8 文献标识码： A 文章编号：1009-5896(2012)10-2348-05 DOI: 10.3724/SP.J.1146.2012.00317 A Laplacian-Cauchy Mixture Model for Improved Correlation Noise Modeling in Distributed Video Coding Yang Chun-ling Xie Xiao-lan (School of Electronic and Information Engineering, South China University of Technology, Guangzhou 510640, China) Abstract: A novel Laplacian-Cauchy Mixture Distribution (LCMD) model is proposed to characterize the distribution of correlation noise between the original Wyner-Ziv frame and the side information in transform domain distributed video coding schemes. Observing the deviations of traditional Lapalcian modeled distribution on large and small residual coefficients, to reduce the deviations, a LCMD model is proposed by modeling small Discrete Cosine Transform (DCT) coefficients as improved Laplacian distributed while modeling large ones as Cauchy distributed. Two solutions to find parameters of LCMD model are also proposed. Real video experiments demonstrate the improvements of LCMD in terms of both the Rate Distortion (RD) performance and computation complexity. Key words: Distributed Video Coding (DVC); Laplacian-Ca