基于小波变换的低通系数差分的图像压缩算法

Difference-based image compression algorithm for wavelet transform with low-pass coefficients

提出了在SPIHT算法中引入差分处理、对低频域小波系数根据其分布特性进行分析的方法．图像经过小波分解后，对生成的小波系数进行差分处理．对5阶分解后的小波系数最顶层采用分块分组的方法，用小波块的质心代替小波系数实现滤波，每组小波块做差分处理，从而减小低频域小波系数间的冗余．在此基础上，通过统计和推导，改进差分量化的模式，得到其优化算法．实验表明，本算法比标准SPIHT的算法具有更高的压缩比，压缩性能能够提高0．01～0．02dB．

Differencial mode being used in set partrtioning in hierarchical trees（SPIHT） algorithm, the analysis of wavelet coefficients with low-pass was made, according their distributing characters in the wavelet-based regions. The coefficients were differentiated after wavelet transform. The top layer coefficients after 5 level decomposed was made to block and group, using the barycenter to replace the coefficients and doing difference in every block, reducing the distribution of redundancy among the wavelet coefficients. Based on the algorithm, improving the mode of difference disposal, optimizing algorithm through the statistics and deduction. The experiments show that the algorithm after optimizing get the favorable compressing capability than that of standard of SPIHT, the quality of compression has improved by 0.01 -0.02 dB.