基于遗传算法的正交对称多小波构造  

Construction of Orthogonal Symmetric Multi-wavelets Based on Genetic Algorithm

作　　者：曾朝晖[1] 刘亚军[2,3] 

机构地区：[1]湘潭大学信息工程学院,湘潭411105 [2]天津大学管理学院,天津300072 [3]湘潭大学商学院,湘潭411105

出　　处：《计算机工程》2010年第11期209-211,共3页Computer Engineering

基　　金：湖南省教育厅一般基金资助项目(09C963)

摘　　要：在多小波参数化的基础上,利用遗传算法的全局优化能力,将带参数的正则指数函数作为遗传算法的适应度函数,构造出正则指数较高的多小波。与其他具有相等支撑长度、消失矩阶数及滤波器结构的多小波相比,该多小波的正则指数更优。对光滑和纹理图像的压缩和去噪的实验结果表明,该多小波的图像处理能力更强、效果更好。On the basis of multi-wavelets parameterization, multi-wavelets with higher regularity are constructed by using total optimal ability of Genetic Algorithm（GA）, in which regular exponential function with parameters is used as fitness function. Contrasting with other multi-wavelets of the same supporting length, vanishing moment and multi-filter bank structure, the multi-wavelet is better in regularity. Compressing and de-noising of smooth image and texture image are raked. Experimental results show that the multi-wavelet has better processing ability and performance.

关 键 词：多小波 遗传算法 正则性 图像压缩 图像去噪 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]