一种减小静止图像压缩效应的后处理算法(英文)  

Novel Post-processing Algorithm to Reduce Still Image Compression Artifacts

作　　者：孙文方[1] 赵亦工[1] 朱红[1] 

机构地区：[1]西安电子科技大学电子工程学院模式识别与智能控制研究所,西安710071

出　　处：《光子学报》2009年第2期430-434,共5页Acta Photonica Sinica

基　　金：Supported by National Natural Science Foundation of China(60572151)

摘　　要：提出了一种静止图像压缩的后处理算法.执行了基于Contourlet变换的图像插值迭代方法来恢复图像的高频通带信息,然后该信息和解压图像的高频信息相融合以进一步保护重要的小波系数.在低码率情况下,该算法有助于减小压缩效应并且提高解压图像沿边缘的正则性.在JPEG2000和SPIHT压缩系统中证实了该方法的性能.仅仅通过一次迭代,后处理图像的峰值信噪比相比于解压图像可提高0.05～0.2dB.当迭代次数增加时,振铃效应将进一步降低.A post-processing algorithm of still image compression was proposed. Iterative contourlet-based image interpolation was implemented to revive the highpass information, which was then fused with the highpass information of the decompressed image to further preserve the significant wavelet coefficients. The algorithm helped to reduce compression artifacts and improve regularity along edges of the decompressed image at low bitrates especially. The ability of the proposed method was demonstrated under JPEG2 000 and SPIHT compression systems. Only for once iteration, the PSNR （peak signal to-noise ratio） of the post-processed image may improve 0.05-0.2 dB than that of the decompressed image. Ringing artifacts decrease with the increase of the number iteration.

关 键 词：静止图像压缩 小波变换 CONTOURLET变换 图像插值 

分 类 号：TP919.81[自动化与计算机技术]