基于提升小波和Context模型的图象无失真编码方案

文章讨论了提升小波变换的基本原理。并根据提升小波系数的特点,结合基于Context模型的熵编码,提出了一种新的无失真图象编码方案。该方案中,将提升小波系数的编码分为三部分考虑,即反映非零系数位置的SignificanceMap、反映非零系数幅值大小的MostSignificanceBits(MSB)符号流和去除了MSB的剩余二进制符号流。因为Signifi-canceMap和MSB符号流中存在较强相关性,提出了简单有效的Context模型对它们进行熵编码。而对去除了MSB的剩余二进制符号流,由于其相关性较弱,直接使用自适应算术编码。经过实验对比证明了该方案在无失真图象编码中的有效性。

Lossless Image Coding Scheme Based on Lifting Wavelet and Context Modeling

The principle of lifting scheme is described.According to the feature of lifting wavelet coefficients,this paper presents a novel lossless image coding system by combining integral lifting wavelet with context modeling.In entropy coding step,wavelet coefficients are divided into three parts,which are Significance Map,Most Significance Bits(MSB)and residual binary bits.The Significance Map represents the location of non-zero coefficients.At the same time ,the MSB represents the extent of non-zero coefficients.Because of the high correlation in the Significance Map and the MSB,simple and effective context model is presented for entropy coding.Because of the low correlation in the residual binary bits,direct arithmetic coding is used.Experiment results illustrate that this method is an effective approach for lossless image coding.