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。

关键词: 多小波, 遗传算法, 正则性, 图像压缩, 图像去噪

Abstract: 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 taked. Experimental results show that the multi-wavelet has better processing ability and performance.

Key words: multi-wavelets, Genetic Algorithm(GA), regularity, image compressing, image de-noising

中图分类号: 

• TP311