第14卷第2期 智能系统学报 Vol.14 No.2 2019年3月 CAAI Transactions on Intelligent Systems Mar.2019 D0:10.11992/tis.201712035 网络出版地址:http:/kns.cnki.net/kcms/detail/23.1538.TP.20180508.1636.002.html 一种融合邻域信息的模糊C-均值图像分割算法 狄岚,刘海涛,何锐波 (江南大学数字媒体学院,江苏无锡214122) 摘要:模糊C-均值算法(fuzzy C-means,FCM)对图像噪声敏感,只考虑了图像数值信息而忽略了邻域空间信 息,造成最终的图像分割结果不精确。为了克服FCM存在的问题,将图像局部信息与非局部信息融入到多测 度模型中,扩充了原本聚类的单一测度。另外将先验概率引入隶属度矩阵中,使得每次迭代前,隶属度矩阵中 像素点的邻域信息都被充分考虑,最后添加一个邻域隶属度惩罚项修正聚类结果。实验证明:该算法对噪声鲁 棒性强,能够获得较为理想的图像分割效果。 关键词:模糊C均值;图像分割:空间信息;局部信息;非局部信息;多测度模型;邻域隶属度:惩罚项 中图分类号:TP3914文献标志码:A文章编号:1673-4785(2019)02-0273-08 中文引用格式:狄岚,刘海涛,何锐波.一种融合邻域信息的模糊C-均值图像分割算法J.智能系统学报,2019,14(2): 273-280. 英文引用格式:DILan,LIU Haitao,,HERuibo.Fuzzy C-means image segmentation algorithm incorporating neighborhood inform- ation/JI.CAAI transactions on intelligent systems,2019,14(2):273-280. Fuzzy C-means image segmentation algorithm incorporating neighborhood information DI Lan,LIU Haitao,HE Ruibo (College of Digital Media,Jiangnan University,Wuxi 214122,China) Abstract:The fuzzy C-means algorithm(FCM)is sensitive to image noise;in addition,it only considers the image nu- merical information and ignores the neighborhood spatial information,resulting in inaccurate final image segmentation result.To overcome this drawback,an FCM image segmentation algorithm is proposed in which the local information and non-local information of the image are integrated into a multidimensional model,which extends the original single dimension of clustering.In addition,a prior probability is introduced into the membership matrix,so that the neighbor- hood information of the pixel in the membership matrix is fully considered before each iteration,and then a neighbor- hood membership penalty is added to correct the clustering result.Finally,a penalty of neighborhood membership de- gree is used to modify the clustering results.Experimental results demonstrate that the algorithm is robust against noise and achieves an ideal image segmentation effect. Keywords:fuzzy C-means;image segmentation;spatial information;local information;non-local information;multidi- mensional model;neighborhood membership degree;penalty term 图像分割在图像识别和计算机视觉中是关键 图像分割成相应结构连贯的元素。根据图像中像 的预处理过程,很多算法已经被提出并应用在对 素的灰度值、纹理、颜色等将图像分为若干个互 象分割以及特征提取中。然而,鲁棒性强和高效 不相干的区域,每个区域内部均有其相似性,不 的图像分割算法的设计还是一个非常具有挑战性 同的区域又互有差异。图像分割可以认为是基于 的研究课题。它的目标是将相似和邻近的像素以 同质性或异质性准则将一幅图像划分为若干有意 义的子区域的过程"。 收稿日期:2017-12-25.网络出版日期:2018-05-14 通信作者:刘海涛.E-mail:382419872@qq.com. 由于待分割模型不尽相同,因此不同模型所DOI: 10.11992/tis.201712035 网络出版地址: http://kns.cnki.net/kcms/detail/23.1538.TP.20180508.1636.002.html 一种融合邻域信息的模糊 C-均值图像分割算法 狄岚,刘海涛,何锐波 (江南大学 数字媒体学院,江苏 无锡 214122) 摘 要:模糊 C-均值算法 (fuzzy C-means,FCM) 对图像噪声敏感,只考虑了图像数值信息而忽略了邻域空间信 息,造成最终的图像分割结果不精确。为了克服 FCM 存在的问题,将图像局部信息与非局部信息融入到多测 度模型中,扩充了原本聚类的单一测度。另外将先验概率引入隶属度矩阵中,使得每次迭代前,隶属度矩阵中 像素点的邻域信息都被充分考虑,最后添加一个邻域隶属度惩罚项修正聚类结果。实验证明:该算法对噪声鲁 棒性强,能够获得较为理想的图像分割效果。 关键词:模糊 C-均值;图像分割;空间信息;局部信息;非局部信息;多测度模型;邻域隶属度;惩罚项 中图分类号:TP391.4 文献标志码:A 文章编号:1673−4785(2019)02−0273−08 中文引用格式:狄岚, 刘海涛, 何锐波. 一种融合邻域信息的模糊 C-均值图像分割算法 [J]. 智能系统学报, 2019, 14(2): 273–280. 英文引用格式:DI Lan, LIU Haitao, HE Ruibo. Fuzzy C-means image segmentation algorithm incorporating neighborhood information[J]. CAAI transactions on intelligent systems, 2019, 14(2): 273–280. Fuzzy C-means image segmentation algorithm incorporating neighborhood information DI Lan,LIU Haitao,HE Ruibo (College of Digital Media, Jiangnan University, Wuxi 214122, China) Abstract: The fuzzy C-means algorithm (FCM) is sensitive to image noise; in addition, it only considers the image numerical information and ignores the neighborhood spatial information, resulting in inaccurate final image segmentation result. To overcome this drawback, an FCM image segmentation algorithm is proposed in which the local information and non-local information of the image are integrated into a multidimensional model, which extends the original single dimension of clustering. In addition, a prior probability is introduced into the membership matrix, so that the neighborhood information of the pixel in the membership matrix is fully considered before each iteration, and then a neighborhood membership penalty is added to correct the clustering result. Finally, a penalty of neighborhood membership degree is used to modify the clustering results. Experimental results demonstrate that the algorithm is robust against noise and achieves an ideal image segmentation effect. Keywords: fuzzy C-means; image segmentation; spatial information; local information; non-local information; multidimensional model; neighborhood membership degree; penalty term 图像分割在图像识别和计算机视觉中是关键 的预处理过程,很多算法已经被提出并应用在对 象分割以及特征提取中。然而,鲁棒性强和高效 的图像分割算法的设计还是一个非常具有挑战性 的研究课题。它的目标是将相似和邻近的像素以 图像分割成相应结构连贯的元素。根据图像中像 素的灰度值、纹理、颜色等将图像分为若干个互 不相干的区域,每个区域内部均有其相似性,不 同的区域又互有差异。图像分割可以认为是基于 同质性或异质性准则将一幅图像划分为若干有意 义的子区域的过程[1]。 由于待分割模型不尽相同,因此不同模型所 收稿日期:2017−12−25. 网络出版日期:2018−05−14. 通信作者:刘海涛. E-mail:382419872@qq.com. 第 14 卷第 2 期 智 能 系 统 学 报 Vol.14 No.2 2019 年 3 月 CAAI Transactions on Intelligent Systems Mar. 2019