基于Chan-Vese模型的甲状腺结节超声图像分割方法  

Thyroid Nodule Ultrasound Image Segmentation Method Based on Chan-Vese Model

作　　者：郑斌 徐峰 郭进祥 刘立波 ZHENG Bin;XU Feng;GUO Jinxiang;LIU Libo(Ningxia University,Yinchuan 750021,Ningxia)

机构地区：[1]宁夏大学信息工程学院,宁夏银川750021

出　　处：《电脑与电信》2018年第6期4-7,共4页Computer & Telecommunication

基　　金：宁夏自然科学基金项目;项目编号:NZ17010;2016宁夏回族自治区研究生教育创新计划;项目编号:YKC201604;国家自然科学基金项目;项目编号:61751215

摘　　要：针对使用Chan-Vese模型分割甲状腺结节时出现的不准确分割现象,提出结合形态学闭运算的超声图像甲状腺结节分割方法。首先,对超声甲状腺图像做预处理,包含图像去噪、增强等操作;然后,做闭运算以及图像归一化;最后,使用Chan-Vese方法分割甲状腺结节区域。实验验证结果表明,结合闭运算的分割可以有效分割出甲状腺结节,并为分割部分甲状腺结节提供了一种有效的分割方法。With regard to the presence of inaccurate segmentation phenomenon in the segmentation of thyroid nodules using Chan-Vese model,an ultrasound image thyroid nodule segmentation method is proposed in combining with morphological closure.First of all,the ultrasound thyroid image is pre-processed,including image denoising,enhancement and other operations;then the closing operation and image normalization are carried on;finally,the Chan-Vese method is used to segment the thyroid nodule region.The experimental verification results show that the segmentation combined with closed operations can effectively segment thyroid nodules and provide an effective method for the segmentation of some thyroid nodules.

关 键 词：CHAN-VESE模型 闭运算 结节分割 超声图像 甲状腺 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]