|
王卫国 郭宝龙 （西安电子科技大学机电工程学院，西安 710071）摘 要 随着互联网的普及和图象应用范围的不断扩大,对图象的编码提出了新的要求,即不仅要求具有高的压缩比,还要求有许多新的功能,如渐进编解码、从有损压缩到无损压缩等。嵌入式零树小波编码较好地实现了这一思想,因此奠定了它在图象编码中的地位。近年来，在嵌入式零树小波编码(EZW)算法的基础上出现了许多新的改进算法，如多级树集合分裂算法(SPIHT),集合分裂嵌入块编码(SPECK),可逆的嵌入小波压缩法(CREW)等.本文对这些算法从原理到性能进行了比较和讨论,说明了嵌入式图象编码的研究方向。关 键 词 图象编码 嵌入式 零树 小波变换On Embedded Zerotree Wavelets Coding and other Improved AlgorithmsWANG Wei-guo, GUO Bao-long(School of Mechano-Electronic Engineering,Xidian Univ.,Xi’an 710071)Abstract With the extensive application of internet and image,some new requirements on image coding,such as high compression rate ,pregressive codec,and compression from lossy to lossless ,are to be satisfied.These functions can be performed well by EZW(Embedded Zerotree Wavelets) coding.On the bases of EZW,many newly improved algorithms have been developed in recent years.They can illustrated by algorithms like SPIHT(Set Partitioning in Hierarchical Trees),SPECK(Set Partitioned Embedded block coder),In this paper,the writer discusses the principles and performances of these algorithms,thus explains the research tendency in the area of embedded image coding.Keywords Image coding,Embedded,Zerotree,Wavelet transform0. 引言在基于小波变换的图象压缩方案中,嵌入式零树小波 EZW(Embedded Zerotree Wavelets)[1]编码很好地利用小波系数的特性使得输出的码流具有嵌入特性。它的重要性排序和分级量化的思想被许多编码算法所采用。近年来,在对EZW改进的基础上，提出了许多新的性能更好的算法,如多级树集合分裂算法(SPIHT :Set Partitioning In Hierarchical Trees)[2],集合分裂嵌入块编码(SPECK：Set Partitioned Embedded bloCK coder),可逆嵌入小波压缩算法(CREW:Compression with Reversible Embedded Wavelets)[3] 。本文对这些算法进行了原理分析、性能比较,说明了嵌入式小波图象编码的研究方向。（完）
|
