摘要
为实现脑肿瘤图像的精确分割,提出一种SOM-FCM的脑肿瘤图像分割算法。对分块后图像进行模极大值边缘检测,将图像块分成边缘类和非边缘类,利用SOM神经网络对非边缘类聚类;将学习到的码本与FCM算法结合,对边缘类图像进行聚类,将2种聚类结果相结合,最终实现脑肿瘤图像的精确分割。实验结果表明,本算法分割精度优于现有2种算法,为脑肿瘤图像分割提供了一种有效方案。
In order to achieve accurate segmentation of brain tumor image,an adaptive brain tumor image segmentation algorithm is presented based on SOM neural network and FCM technique.Firstly,the segmented image is divided into small sub-blocks of the same size,the sub-block vectors are classified into edge pattern and non-edge pattern by using an edge detection algorithm.Then the non-edge pattern vectors are clustered by using SOM-based technique,the codebook is learned as the initial clustering centers of the FCM technique,which is used to process the edge pattern.The experimental results show that the segmentation accuracy is superior to the existing two algorithms,which provides an effective scheme for brain tumor image segmentation.
作者
申亮亮
首照宇
SHP＇N Liangliang, SHOU Zhaoyu(School of Information and Communication,Guilin University of Electronic Technology,Guilin 54 1004,Chin)
出处
《桂林电子科技大学学报》
2018年第1期50-54,共5页
Journal of Guilin University of Electronic Technology
基金
国家自然科学基金(61362021
61661017)
广西科技研究与技术开发计划(桂科能1598025-21)
广西自然科学基金(2013GXNSFDA019030
2014GXNSFDA118035
2016GXNSFAA380149)
认知无线电教育部重点实验室基金(CRKL150103
2011KF11)
关键词
脑肿瘤分割
神经网络
边缘检测
码本
brain tumor segmentation
neural network
edge detection
codcbook
