Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (22): 191-193.DOI: 10.3778/j.issn.1002-8331.2009.22.061

• 工程与应用 • Previous Articles     Next Articles

Optimization of multi-thresholding segmentation in head CT images

LIU Si-wei1,LI Bin1,TIAN Lian-fang1,CHEN Ping2   

  1. 1.College of Automation Science and Engineering,South China University of Technology,Guangzhou 510640,China
    2.Dept. of Nuclear Medicine,The First Affiliated Hospital of Guangzhou Medical College,Guangzhou 510120,China
  • Received:2008-04-22 Revised:2008-07-11 Online:2009-08-01 Published:2009-08-01
  • Contact: LIU Si-wei

头部CT图像多阈值分割的优化实现

刘思伟1,李 彬1,田联房1,陈 萍2   

  1. 1.华南理工大学 自动化科学与工程学院,广州 510640
    2.广州医学院 第一附属医院 核医学科,广州 510120
  • 通讯作者: 刘思伟

Abstract: Thresholding segmentation is one of the methods of medical image pre-processing.An appropriate method can reduce the workload of the subsequent processing.However,the segmentation has low efficiency with applying only one traditional method each time.This article introduces a multi-thresholding segmentation method using Otsu with genetic algorithm.The method is applied in the head CT image segmentation,and compared with the overall searching algorithm.The result shows that it keeps the accuracy of overall searching algorithm and increases the searching speed.Therefore,the efficiency of the segmentation is enhanced dramatically.

Key words: multi-thresholding segmentation, genetic algorithm, head CT images, Otsu

摘要: 阈值分割是医学图像预处理方法的一种,合适的分割方法能减少后续图像处理的数据。单独使用传统的阈值分割方法往往效率很低。提出一种基于遗传算法的图像多阈值分割方法。通过设计最大类间方差法与遗传算法相结合的算法,对头部CT图像进行分割,并与遍历算法作比较。实验结果表明,此方法不仅保留遍历算法的精度,并且快速得到最优阈值,明显提高了分割的效率。

关键词: 多阈值分割, 遗传算法, 头部CT图像, 最大类间方差法