摘要
以截集模糊C均值聚类(Sectional Set Fuzzy C-means algorithm:SSFCM)算法为基础,提出一种自适应遗传算法(Adaptive Genetic Algorithm)改进截集FCM算法。传统FCM算法中一般使用一维直方图初始化方法,使初始化与聚类算法相分离,没有形成整体,而且同一幅一维直方图可能对应不同的原始图像。引入自适应遗传算法,与截集FCM算法有机结合,用遗传算法解决初始化问题的同时,以遗传算法的寻优性能来指导聚类。实验表明,该算法效率较传统FCM算法和未改进截集FCM算法有很大的提高,同时能够保持较好的分割效果和质量。
An improved method based on Sectional Set Fuzzy C-means (SSFCM) and Adaptive Genetic Algorithm (AGA) algorithm applied in image segmentation is presented. In the traditional FCM method, Histogram is used for initializing, but it can not be connected with the FCM as a whole and can not match to the exclusive picture. So, AGA is used to optimize the SSFCM, initialize and direct clustering. The experimental results prove rate that the new algorithm's efficiency is higher than both traditional FCM and unimproved SSFCM, while maintains a good segmentation effect.
出处
《西安科技大学学报》
CAS
北大核心
2006年第1期85-88,共4页
Journal of Xi’an University of Science and Technology
关键词
图像分割
截集模糊C均值聚类算法
自适应遗传算法
image segmentation
sectional set fuzzy C-means algorithm
adaptive genetic algorithm