改进的图像分割遗传K-均值聚类算法

An improved genetic K-means clustering algorithm based on image segmentation

针对图像分割,提出了一种改进的遗传K-均值聚类算法。合理选取聚类的特征向量并对各特征分量确定不同权值进行调整;通过引入自适应算法,对传统遗传算法的选择及变异操作进行改进,提高了算法的收敛速度;确定与染色体编码相关的隶属矩阵可有效地减少运算时间。实验结果表明,改进后的遗传K-均值聚类算法是行之有效的。

An improved genetic K-means clustering algorithm was described based on image segmentation. The feature vector of the pixel was properly chosen and the weight factors of the feature vector were adjusted. The selection of conventional genetic algorithm and the mutation operations were improved by the introduction of adaptive algorithm, which enhanced the speed of convergence. Computing time was reduced by determining the membership matrix related to the code of chromosomes. The results of the experiments demonstrate that the proposed algorithm is more effective than the traditional genetic K-means algorithm.