一类整数小波变换的优化设计及其图像编码应用

Optimization design of a class of integer wavelet transforms and their application to image coding

整数小波变换已在嵌入式图像编码领域取得了极大的成功,一种实现整数小波变换的方法是提升.给出一种基于提升的参数化构造法,构造一类新的整数小波变换,它们仅用一个自由参数表示.随意调整该参数,就可构造出不同的整数小波变换.作为构造示例,构造了5种整数小波变换,它们的提升滤波器系数全为二进制分数(形如k/2n,k,n∈Z的数).实验表明,其中的两种变换与当前几种广为应用的整数小波变换相比,具有优异的有损和无损图像压缩性能.

The integer wavelet transform （IWT） has proved to be particularly successful in the area ot embedded image coding. One of the possible methods to realize the IWT is the lifting scheme. A parameterization technique of IWT was presented based on the lifting scheme and a new class of IWT was parameterized by using one free parameter only. Different IWT could easily be obtained by adjusting the free parameter. As an illustration of this method, five IWTs with their lifting filters all having binary coefficients （numbers of the form k/2^n,n∈Z） were constructed. The experiments showed that, compared to several other widely-used IWTs, two of these transforms possessed more superior compression performance for both lossy and lossless image coding.