基于遗传算法与模糊聚类相结合的小波变换矢量量化方法

A wavelet transformation vector quantization method based on the combination of genetic and fuzzy c-mean clustering algorithms

提出一种有效的图像压缩方法,利用小波变换对图像进行多分辨率分解,对小波系数进行矢量量化(VQ)编码。使用遗传算法(GA)与模糊c均值聚类(FCM)算法相结合的方法来设计码书,有效地克服了FCM算法容易陷入局部最优且对初始值敏感的缺点。实验结果表明,该算法可较大提高图像重构质量。

An effective image compress method is put forward in this paper.Using wavelet transformation to decompose im-ages,wavelet coefficients are vector quantized.Codebook of vector quantization is designed through fuzzy c-mean clustering and genetic algorithm.This algorithm overcomes the shortcomings of FCM effectively.Experimental results indicate that this algorithm improves the quality of reconstructed image.