极化码在压缩图像传输系统中的性能改进方法  

THE METHOD TO IMPROVE THE PERFORMANCE OF POLAR CODE IN THE COMPRESSED IMAGE TRANSMISSION SYSTEM

作　　者：沈周青 尚俊娜[1] Shen Zhouqing;Shang Junna(College of Telecommunication Engineering,Hangzhou Dianzi University,Hangzhou 310018,Zhejiang,China)

机构地区：[1]杭州电子科技大学通信工程学院,浙江杭州310018

出　　处：《计算机应用与软件》2018年第3期145-150,共6页Computer Applications and Software

基　　金：浙江省自然科学基金青年基金项目(LQ13F010010);浙江省重点科技创新团队项目(2013TD03);浙江省"电子科学与技术"重中之重学科开放基金项目(GK13020320003/004)

摘　　要：为了解决图像经压缩去冗余编码后易受干扰的问题,提出一种Polar码在嵌入式零树小波(EZW)压缩图像传输系统中的性能改进方法。在总码率一定的情况下,通过简化组合最优算法,完成对原始压缩码流的分段以及码率分配。根据Polar码存在的不等错误属性和同一段内码流重要性不同,完成不等错误保护。通过仿真,提出的改进方法有效改善了标准Polar码在低信噪比和低密度奇偶校验码(LDPC)码在高信噪比条件下性能表现差的问题,其中,当信噪比为1 d B时,相比于传统Polar码,改进方法重构图像的峰值信噪比(PSNR)值提高了5 d B,相对于LDPC码,对应的PSNR值提高了4 d B。经EZW编码后的码流以重要性进行排序,并不依赖图像信源,因此该方法具有很好的通用性。An approach which improves the Polar codes performance in image compression system by embedded zero tree wavelet encoding is proposed to solve the problem that image is susceptible to interference after compression and redundant coding.When the total bit rate is constant,the segmentation and rate allocation of original compressed bit stream are completed by simplifying the optimal combination algorithm.The unequal error protection is completed based on different error attributes of polar codes and different significance of bit stream within the same segmentation.Through simulation,the proposed method improves the low performance of standard Polar code in low signal to noise ratio(SNR)and the low-density parity check(LDPC)code under high SNR.When the SNR is 1Db,the peak signal-to-noise ratio(PSNR)of the reconstructed image is improved by 5 dB compared to the conventional Polar code,and the corresponding PSNR value is increased by 4 dB relative to the LDPC code.The EZW encoded stream is sorted by significance with no reliance on image source,which means this method has good generality.

关 键 词：不等错误保护 通用性 Polar码 EZW压缩编码 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]