华南理工大学学报(自然科学版) ›› 2008, Vol. 36 ›› Issue (1): 72-75.

• 电子、通信与自动控制 • 上一篇    下一篇

胸部CT中肺实质的自动分割与计算机辅助诊断

郭圣文陈坚曾庆思2   

  1. 1. 华南理工大学 生物医学工程系, 广东 广州 510640; 2. 广州医学院附属第一医院 放射科, 广东 广州 510120
  • 收稿日期:2007-03-14 修回日期:2007-06-27 出版日期:2008-01-25 发布日期:2008-01-25
  • 通信作者: 郭圣文(1971-),男,博士,讲师,主要从事医学图像处理与计算机辅助诊断方面的研究. E-mail:shwguo@scut.edu.cn.
  • 作者简介:郭圣文(1971-),男,博士,讲师,主要从事医学图像处理与计算机辅助诊断方面的研究.

Automatic Segmentation of Pulmonary Parenchyma in Thoracic CT and Computer-Aided Diagnosis

Guo Sheng-wen1  Chen Jian1  Zeng Qing-si2   

  1. 1. Department of Biomedical Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China; 2. Department of Radiology, the First Affiliated Hospital of Guangzhou Medical College, Guangzhou 510120, Guangdong, China
  • Received:2007-03-14 Revised:2007-06-27 Online:2008-01-25 Published:2008-01-25
  • Contact: 郭圣文(1971-),男,博士,讲师,主要从事医学图像处理与计算机辅助诊断方面的研究. E-mail:shwguo@scut.edu.cn.
  • About author:郭圣文(1971-),男,博士,讲师,主要从事医学图像处理与计算机辅助诊断方面的研究.

摘要: 为解决肺气肿的定量分析与辅助诊断问题,根据肺部高分辨率CT图像的特点,应用自动阈值分割与轮廓跟踪方法提取肺部实质区域,并应用基于密度分布的肺气肿量化诊断标准确定病变区域与程度.实验证明,该方法能准确、有效地对肺部区域实行全自动分割,并对病变区域进行统计分析,最终实现肺气肿的量化分析与准确诊断.

关键词: 高分辨率CT 医学图像, 肺部区域, 分割, 辅助诊断

Abstract:

In order to implement the quantitative analysis and aided diagnosis of emphysema, an aaaputive thresholdsegmenting and contour-tracing method was proposed to automatically extract the information of pulmonary parenchyma from high-resolution CT images, and the region and severity of emphysema were then determined according to the quantitative analysis criterion based on density distribution. Experimental results show that the proposed method can segment the pulmonary parenchyma automatically with accurate and effective results, and can statistically analyze the pathological change regions, thus implementing the quantitative analysis and exact diagnosis of emphysema.

Key words: high resolution CT, medical image, lung region, segmentation, aided diagnosis