一种基于遗传算法的图象矢量量化方法

描述了一种基于遗传算法的图象矢量量化方法．遗传算法基于自然进化原理可得到全局优化结果，把遗传算法用于码本设计，其运算简单，而且可避免传统方法如Ｋ－均值技术等所带来的局部优化局限．同时，由于其结构的规则性和并行性，较适于ＶＬＳＩ实现．文中主要讨论了遗传算法用于图象矢量量化的评价函数及操作过程，并与传统算法作了比较，给出实验结果．

Image VQ Coding Using Genetic Algorithm

This poper presents a new approach to image compression using vector quantitation(VQ)technique and genetic algorithm(GA).GA is an optimal algorithm which comes from the principle of natural process.By using GA for image VQ codebook design,It is found that GA is more robust in locating the global optimum and is computationally simpler than other algorithms such as k-means.Because of its regularity and high parallelism in architecture,this method is also suitable for VLSI implementation. The objective function and operation process of GA used for image VQ are discussed in detail its perfomance is evaluated through some experiment results.