摘要
为了提高诊断的准确率和效率,提出了一种新的连通域标记算法,同时对医学图像中感兴趣区域进行连通域标记和区域特征提取。该算法先对读入二值图进行边界提取,再对边界进行跟踪和标记,利用图像重构的方法对边界进行区域填充,并将属于同一连通域的边界进行归类,即重新排列标记号,最后对连通区域的形态特征进行提取。实验证明,该算法不但能正确标记任意复杂形状的连通域,运行速度较快,而且对连通区域进行了特征提取,现已应用到医学图像处理的多个方面,为下一步的图像处理奠定了更好的基础。
In order to enhance the accuracy and efficiency of disease diagnosis,this paper presented a new connected components labeling algorithm,which could not only label regions of interest(ROIs) in medical images but also obtain features from ROIs.It began with detecting boundaries of the input boundary image,then applied boundary tracing and labeling,filled next boundaries by using image reconstruction and classified by judging whether they belonged to the same connected component,which meant rearranging the marks.Finally,features of connected components would be obtained.The present results indicate that our scheme is robust in labeling connected components of any shape with good speed,meanwhile,it can obtain features of connected components.So far it has been applied to many fields in medical image processing,which definitely provides a better basis for the following image processing.
出处
《计算机应用研究》
CSCD
北大核心
2010年第11期4338-4340,共3页
Application Research of Computers
基金
中国博士后基金资助项目(20090450866)
国家教育部博士点基金资助项目(200805610018)
广东省教育部产学研结合项目(2009B090300057)
广东省自然科学基金资助项目(8451064101000631)
广州市番禺区科技攻关项目(2009-Z-108-1)
关键词
医学图像处理
连通区域标记
边界提取
图像重构
特征提取
medical image processing
connected components labeling
boundary detection
image reconstruction
features extraction