基于复原理论的小波域压缩图像后处理算法.pdfVIP

74 2007 ,43(29) Computer Engineeri咆 and Applica1ions 计算机工程与应用 基于复原理论的小波域压缩图像后处理算法 肖志云 1 ，文伟 2 ，彭思龙 2 XIAO Zhi-yun1 ,WEN Wei2 ,PENG Si-long2 1.内蒙古工业大学信息工程学院，呼和浩特 010051 2 中国科学院自动化研究所国家专用集成电路设计工程技术研究中心，北京 100080 I.College of Information Engineering ,Inner Mongolia University of Teehonolgy ,Hohhot 010051 ,China 2.National ASIC Design Engineering Center , Institute of Automation ,Chinese Ac- ademy of Sciences , Beijing 100080 ,China E-lI1ail: xiaozhiyun@imu t. XIAO Zhi-yun. WEN Wei.PENG Si 斗ong.Postprocessing algorithm for DWT -based compressed image based on image restore theory.Computer Engineering and Applications ,2007 .43( 29): 74-77. Abstract: Recently , image compression algorithms based on Discrete Wavelet Transform( DWT) ell1erge in plentv. But at low bit- rate ,c-oded images suffer from serious quantization noise ,which will degrade the quality of coded images.ln this paper ,a new postprocessing algorithm based on image restore theoηis proposed.Based on the proposed postprocessing lI1odel , constrain刷i Least-squares lI1ethod is used to restore the coded ill1age.Guidable optimal parall1eters selection is given based on nU ll1erical sim ulation trials.Results show that the proposed algorithm can improve the visual effects of the coded images as well as the PSNR gain.Coll1pared with other existent algorithms. the proposed algorithm shows its advantages especially at low bit-rate Key words: compressed image postprocessing; least-squares method; DWT; image restoration 摘 要:近来，基于小波变换的各种压缩算法大量出现υ 但在低比特率情况下，由于小泼系数的量化将导致各种量化噪声，严重破 坏了压缩图像的质量。针对这种情况，提出了一种基于复原理论的小波变换压缩图像后处理算法3 在建立的后处理模型上，应用约 束最小二来法求解。通过试验，给出了最优参数的指导值选取c 试验结果表明，该算法可以提高低比特率下压缩图像的视觉效果， 同时 PSNR 也有一定的提高。和其它算法的对比试验表明，该算法在低比特率情况下有一定的优越性。 关键词:压缩图像后处理;最小二乘;小波变换;图像复原 文章编号:1002-8331(2007)29-0074-04 文献标识码:A 中罔分类号:TP39 1 态图像压缩标准和 MPEG 动态压缩标准都采用了 BDCT。基于 1 寻|言