结合非局部信息的模糊聚类脑MR图像分割
发布时间：2018-08-17 19:09
【摘要】：为降低噪声以及偏移场的影响,提出一种基于非局部空间信息的FCM模型。引入非局部信息,克服传统的空间信息仅依赖邻域灰度信息,导致精度不高的缺点,使其在降低噪声影响的同时保持细长拓扑结构区域信息;利用多元高斯分布模型对图像灰度分布进行拟合,构造距离函数,降低传统欧式距离导致鲁棒性不足的影响;利用基函数的线性组合对偏移场进行拟合,将偏移场参数化并耦合到FCM框架下,降低灰度不均匀对分割的影响。实验结果表明,该模型可以得到更精确的分割结果。
[Abstract]:In order to reduce the influence of noise and offset field, a FCM model based on nonlocal spatial information is proposed. The nonlocal information is introduced to overcome the disadvantage that the traditional spatial information only depends on the gray level information of the neighborhood, which leads to the low precision, which can reduce the influence of noise and keep the information of the slender topological structure region at the same time. The multivariate Gao Si distribution model is used to fit the gray distribution of the image, and the distance function is constructed to reduce the influence of the traditional Euclidean distance on the robustness, and the linear combination of the basis function is used to fit the offset field. The offset field is parameterized and coupled to the FCM framework to reduce the influence of gray inhomogeneity on segmentation. Experimental results show that the model can obtain more accurate segmentation results.
【作者单位】： 南京信息工程大学数学与统计学院;
【基金】：江苏省自然科学基金项目(BY2014007-04)
【分类号】：R445.2;TP391.41
本文编号：2188610
[Abstract]:In order to reduce the influence of noise and offset field, a FCM model based on nonlocal spatial information is proposed. The nonlocal information is introduced to overcome the disadvantage that the traditional spatial information only depends on the gray level information of the neighborhood, which leads to the low precision, which can reduce the influence of noise and keep the information of the slender topological structure region at the same time. The multivariate Gao Si distribution model is used to fit the gray distribution of the image, and the distance function is constructed to reduce the influence of the traditional Euclidean distance on the robustness, and the linear combination of the basis function is used to fit the offset field. The offset field is parameterized and coupled to the FCM framework to reduce the influence of gray inhomogeneity on segmentation. Experimental results show that the model can obtain more accurate segmentation results.
【作者单位】： 南京信息工程大学数学与统计学院;
【基金】：江苏省自然科学基金项目(BY2014007-04)
【分类号】：R445.2;TP391.41
【相似文献】
相关硕士学位论文 前10条
1 周连波;基于局部信息的复杂网络社团挖掘算法研究[D];燕山大学;2015年
2 于健;带有随机扰动的风场优化调度策略的研究[D];北京交通大学;2016年
3 牛会兰;局部信息方法在功能脑网络建模中的应用研究[D];太原理工大学;2016年
4 贾桂敏;基于物体局部信息的跟踪算法研究[D];天津大学;2008年
5 龙丽君;网络内容监管系统中基于局部信息的语义倾向性识别算法[D];南京理工大学;2004年
6 王聪;基于局部信息交流的命名游戏问题研究与应用[D];大连理工大学;2013年
7 谢德燕;局部信息保持投影及应用[D];西安电子科技大学;2010年
8 杨雨泽;基于整体和局部信息的多经验核学习及其应用研究[D];华东理工大学;2014年
9 夏磊;一种基于局部信息的社会网络聚类算法[D];哈尔滨工程大学;2012年
10 陈书杨;局部信息缺失情况下人脸识别研究[D];辽宁大学;2014年
，本文编号：2188610
本文链接：https://www.wllwen.com/yixuelunwen/fangshe/2188610.html
