核空间自适应抑制式直觉模糊图像分割算法  

Kernel spatial adaptive suppressed intuitionistic fuzzy C-means image segmentation algorithm

作　　者：林洋 兰蓉 LIN Yang;LAN Rong(School of Telecommunications and Information Engineering,Xi ’an University of Posts and Telecommunications,Xi ’an 710121,China;Key Laboratory of Electronic Information Processing with Application in Crime Scene Investigation of Ministry of Public Security,Xi ’an University of Posts and Telecommunications,Xi ’an 710121,China;Shaanxi International Joint Research Center of Wireless Communication and Information Processing,Xi ’an University of Posts and Telecommunications,Xi ’an 710121,China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121 [2]西安邮电大学电子信息现场勘验应用技术公安部重点实验室,陕西西安710121 [3]西安邮电大学陕西省无线通信与信息处理技术国际合作研究中心,陕西西安710121

出　　处：《计算机工程与设计》2019年第8期2353-2360,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61571361、61671377);陕西省教育厅科学研究计划基金项目(16JK1709);西安邮电大学西邮新星团队基金项目(xyt2016-01)

摘　　要：针对直觉模糊C-均值(intuitionistic fuzzy C-means,IFCM)算法未考虑图像像素的空间邻域信息,导致对噪声较为敏感,算法运行效率较低,分割效果较差等问题,提出一种核空间自适应抑制式直觉模糊图像分割算法。以核诱导距离代替欧氏距离计算像素至聚类中心的距离,将局部空间信息融入核空间中;利用“投票模型”将模糊集扩展为直觉模糊集,减少人工参数对实验的影响;根据图像像素和聚类中心之间的分离性自适应生成抑制因子。实验结果表明,该算法对噪声鲁棒性较强,分割精度较高,提升了算法的运行效率。For dealing with issues of the intuitionistic fuzzy C-means (IFCM) algorithm,such as lack of consideration of the spatial neighborhood information of pixels,noise sensitivity,low algorithm operating efficiency and poor segmentation result,a kernel spatial adaptive suppressed intuitionistic fuzzy C-means image segmentation algorithm was proposed.Instead of Euclidean distance,kernel-induced distance was used to measure the distance from pixels to cluster centers for integrating local spatial information into kernel space.The traditional fuzzy C-means set was developed to intuitionistic fuzzy C-means set by the voting model for reducing the influence of artificial parameters on the experiment.According to the separation between the pixels and the cluster centers,the inhibitory factor was adaptively produced.Experimental results show that the proposed algorithm is robust to noise with better segmentation accuracy.The efficiency of the algorithm is also improved.

关 键 词：直觉模糊C-均值 核空间 局部空间信息 投票模型 抑制因子 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]