摘要
针对肝脏分割结果中出现的误分割问题,在分析腹部CT图像序列中肝脏结构特性和成像特点的基础上,充分利用肝脏灰度特性和结构特性的互补信息,提出一种组合三视图分割结果的置信连接肝脏自动分割方法。从各视图的最佳角度提取肝脏轮廓线,将其合并得到最终的肝脏轮廓,有效减少了肝脏与相邻器官灰度相似性造成的误分割现象,进一步提高了肝脏自动分割的准确率。对10套腹部CT数据集的实验结果表明,该方法能够快速、准确的提取肝脏轮廓,为临床肝病诊断和手术计划制定提供有效的个体化信息。
In order to solve the problem of misclassification in liver segmentation process, an automatic confidence connected liver segmentation method with a combination of the liver segmentation results of three views was proposed. Based on the analysis of liver structure and its imaging features on CT images, this method took full advantage of the liver intensity specialty and shape specialty. The final liver boundary is the combination of the contours extract from best angle of each view. The misclassifications caused by the gray similarity of liver and adjacent organs were reduced. The accuracy of automatic liver segmentation was improved. Clinical validation was performed on 10 abdominal CT data-sets. The results show that the proposed method can extract liver contour quickly and accurately. And it is useful for clinical liver diagnosis and surgery planning.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2014年第9期1980-1983,共4页
Journal of System Simulation
基金
国家自然科学基金(61102137
61327001)
河南省科技厅重点科技攻关计划项目(142102310298)
河南省教育厅科学技术研究重点项目基础研究(14A520056)
南阳师范学院专项项目(ZX2013012)
关键词
肝脏分割
CT图像
三视图
置信连接
liver segmentation
CT image
three views
confidence connected
