基于多进制LDPC码的分布式视频编码  

Distributed Video Coding Based on N-nary Low Density Parity Check Code

作　　者：潘旭洲[1] 刘荣科[1] 王潇[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100191

出　　处：《计算机工程》2011年第17期245-247,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60702012);留学回国人员科研启动基金资助项目

摘　　要：对基于多进制低密度奇偶校验(LDPC)码的分布式视频编码性能进行研究。根据多进制信源的特性,利用拉普拉斯分布改进多进制信源的相关模型,并修正多进制LDPC码的快速译码算法FFT-BP,使其适用于分布式视频编码系统。仿真结果表明,在峰值信噪比相同的条件下,与基于二进制LDPC码的系统相比,基于四进制LDPC码的系统码率平均降低0.028 3 bit/pixel,基于八进制LDPC码的系统码率平均降低0.037 0 bit/pixel,相比同类分布式信源编码系统,该系统的译码速度更快、复杂度更低。This paper researches distributed video coding performance based on n-nary Low Density Parity Check（LDPC） code.According to the characteristics of the n-nary sources,Laplacian distribution is used to improve the correlation model of the n-nary sources.FFT-BP algorithm of n-nary LDPC code is modified to be applicable to the system.Simulation results show that,with the same value of Peak Signal to Noise Ratio（PSNR）,the average code rate of distributed video coding using 4-ary LDPC codes is reduced by 0.028 3 bit/pixel compared to the distributed video coding using binary LDPC codes and the average code rate of distributed video coding using 8-ary LDPC codes is reduced by 0.037 0 bit/pixel.And compared with other distributed source coding systems,the decoding speed is higher and the complexity is lower.

关 键 词：分布式视频编码 多进制低密度奇偶校验码 拉普拉斯分布 相关模型 FFT-BP算法 

分 类 号：TN919.8[电子电信—通信与信息系统]