摘要
为提升模糊C均值(fuzzy C-means,FCM)聚类算法抑制噪声能力,提出一种改进的模糊聚类图像分割算法。首先,利用非局部空间信息和局部空间信息构建多维图像信息。其次,引入先验概率保证每次迭代前的隶属度考虑图像空间信息。最后,添加隶属度惩罚项改善聚类分割效果。实验结果表明,改进算法与模糊局部信息C均值(fuzzy local information C-means,FLICM)聚类算法、结合核度量及局部信息的模糊C均值(KWFLICM)聚类算法、结合非局部空间信息的模糊C均值(FCM_NLS)聚类算法、粒子群优化非局部模糊C均值聚类图像分割(PSO-WMNLFCM)算法相比,划分系数和划分熵均有较大改善,划分系数提高到0.9573,划分熵降低到0.0596。
In order to improve the ability of fuzzy C-means(FCM)clustering algorithm to suppress noise,an improved fuzzy clustering image segmentation algorithm was proposed.First,multi-dimensional image information was set up using non-local spatial information and local spatial information.Secondly,a priori probability was introduced to ensure that the membership degree before each iteration takes into account the spatial information of the image.Finally,the membership penalty item was added to improve the clustering segmentation effect.Experimental results show that compared with fuzzy local information C-means(FLICM)clustering algorithm,fuzzy C-means with local information and kernel metric(KWFLICM)clustering algorithm,fuzzy C-means with non-local spatial information(FCM_NLS)clustering algorithm,and non-local fuzzy C-means clustering image segmentation algorithm based on particle swarm optimization(PSO-WMNLFCM),the improved algorithm has greatly improved partition coefficient and partition entropy,the partition coefficient increased to 0.9573,and the partition entropy is reduced to 0.0596.
作者
刘旭东
李云红
屈海涛
苏雪平
谢蓉蓉
LIU Xudong;LI Yunhong;QU Haitao;SU Xueping;XIE Rongrong(School of Electronics and Information,Xi’an Polytechnic University,Xi’an 710048,China;Harbin Product Quality Supervision and Inspection Institute,Harbin 150036,China)
出处
《西安工程大学学报》
CAS
2021年第3期67-73,共7页
Journal of Xi’an Polytechnic University
基金
国家自然科学基金(61902301)
陕西省科技厅青年科学基金(2019JQ-255)
西安市科技局高校人才服务企业项目(2019217114GXRC007CG008-GXYD7.2)。
关键词
图像分割
模糊C均值
空间信息
聚类算法
惩罚项
image segmentation
fuzzy C-means
spatial information
clustering algorithm
penalty term