基于改进布谷鸟优化的模糊聚类图像分割

发布时间：2018-05-03 00:44

本文选题：图像分割 + 改进布谷鸟优化算法　；参考：《计算机科学》2017年06期

【摘要】：模糊C均值聚类算法(FCM)是一种应用非常广泛的聚类算法,但是它受初始聚类中心影响较大,容易陷入局部最优。在标准布谷鸟算法(CS)的基础上提出改进布谷鸟优化算法(ICS),将发现概率P由固定值转变成随迭代次数逐渐减小的变量,这样不仅可以提高搜索种群的质量,而且保证了算法的收敛。因此,可以将改进布谷鸟优化算法用于FCM算法聚类中心生成的过程(ICS_FCM),从而有效地避免FCM陷入局部最优。改进的算法具有良好的聚类效果和运行速度。实现基于改进布谷鸟优化的FCM图像分割,并与基于模拟退火的FCM算法(SA_FCM)进行对比。由实验结果可知,该算法(ICS_FCM)不仅取得了较好的分割效果,效率上也有明显的提高。
[Abstract]:Fuzzy C-means clustering algorithm (FCM) is a very widely used clustering algorithm, but it is influenced by the initial clustering center and is prone to fall into local optimum. Based on the standard Cuckoo algorithm (CSS), an improved Cuckoo optimization algorithm is proposed, in which the probability of finding P is changed from a fixed value to a variable gradually decreasing with the iteration times, which not only improves the quality of the search population, but also improves the quality of the search population. And the convergence of the algorithm is ensured. Therefore, the improved Cuckoo optimization algorithm can be applied to the process of generating clustering centers of FCM algorithm, thus effectively avoiding FCM falling into local optimum. The improved algorithm has good clustering effect and running speed. FCM image segmentation based on improved cuckoo optimization is realized and compared with FCM algorithm based on simulated annealing. The experimental results show that the algorithm can not only achieve a good segmentation effect, but also improve the efficiency of the algorithm.
【作者单位】：南京邮电大学计算机学院;江苏省无线传感网高技术研究重点实验室;
【基金】：国家自然科学基金项目(61300239,61572261) 中国博士后科学基金资助项目(2014M551635) 江苏省博士后科研资助计划项目(1302085B) 江苏省政府留学基金(JS-2014-085)资助
【分类号】：TP18;TP391.41

【相似文献】