摘要
针对传统模糊聚类初值敏感、易陷入局部最优的缺陷,将具有良好勘探和开采能力的人工免疫算法用于模糊聚类的优化并提出了相应的图像分割算法.利用改进的Hausdorff距离提出一种新的抗体浓度评价算子并定义了相应的免疫算子,简化了免疫操作,增强了算法自适应寻优能力.采用最近提出的一种有效性函数作为聚类适应度函数,以人工免疫算法寻优,从而自适应地确定聚类数目与中心,实现自动图像分割.仿真实验表明,该算法可以实现图像的自动高有效性分割.
For addressing prematurity and initial sensitive problems with traditional fuzzy clustering,artificial immune algorithm is utilized for optimizing fuzzy clustering image segment,which has excellent ability on exploration and exploitation.A new method for antibody density estimation is proposed based on improved Hausdorff distance,and corresponding immune operators are defined.A new validity index function is selected as fitness function.The number and centers of clusters are adaptively decided by searching optimization using artificial immune algorithm,which realizes automatic image segment.Simulation results show that proposed algorithm can automatically segment image with high validity.
出处
《控制与决策》
EI
CSCD
北大核心
2010年第11期1679-1683,共5页
Control and Decision
关键词
图像分割
模糊聚类
有效性函数
人工免疫算法
Image segmentation
Fuzzy clustering
Validity index
Artificial immune algorithm
