摘要
为了提高图像分割的运算速度,该文在将传统模糊C均值(FCM)聚类算法应用于图像自动分割的基础上,提出一种改进的快速图像分割算法。将图像从像素空间映射至其对应的灰度直方图特征空间,实现在特征空间进行数据聚类分析以减少聚类样本数量。依据灰度直方图特性,通过曲线拟合方法获得图像的聚类数及初始聚类中心。实验结果表明,在有效分割图像的基础上,该算法的运算迭代次数减少了约10%,运行时间减小了约6%。
In order to improve the speed of image segmentation ,an improved fast image segmentation algorithm is proposed based on the application of conventional fuzzy C-means （ FCM ） clustering algorithm for image automatic segmentation .An image is mapped to the corresponding gray histogram feature space from the pixel space .Data clustering analysis is realized in the feature space and the amount of the clustering sample is reduced .According to the characteristics of the gray histogram ,the clustering number and the initial clustering center are obtained by curve fitting method .The experimental results show that the iterative number is reduced by 10%and the run time is reduced by 6%and effective image segmentation is realized .
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2016年第3期309-314,共6页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金(51175262)
安徽省高校自然科学研究项目(KJ2013B075)
安徽科技学院青年科学基金(ZRC2013338)
安徽科技学院重点建设学科(AKZDXK2015C02)
关键词
模糊聚类
C均值聚类
图像分割
像素空间
灰度直方图
特征空间
曲线拟合方法
聚类数
初始聚类中心
fuzzy clustering
C-means clustering
image segmentation
pixel space
gray histogram
feature space
curve fitting method
clustering number
initial clustering center
