计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (20): 236-239.DOI: 10.3778/j.issn.1002-8331.2009.20.069

• 工程与应用 • 上一篇    下一篇

多分类支持向量机分割彩色癌细胞图像

窦智宙1,平子良2,冯文兵3,王永祥3   

  1. 1.内蒙古师范大学 计算机信息工程学院,呼和浩特 010022
    2.内蒙古师范大学 物理与信息工程学院,呼和浩特 010022
    3.内蒙古医学院 附属医院,呼和浩特 010050
  • 收稿日期:2008-04-18 修回日期:2008-08-04 出版日期:2009-07-11 发布日期:2009-07-11
  • 通讯作者: 窦智宙

Segmentation of multicolor cancer cells with Multi-class Support Vector Machine

DOU Zhi-zhou1,PING Zi-liang2,FENG Wen-bing3,WANG Yong-xiang3   

  1. 1.College of Computer and Information Engineering,Inner Mongolia Normal University,Hohhot 010022,China
    2.College of Physical and Electronic Information,Inner Mongolia Normal University,Hohhot 010022,China
    3.Inner Mongolia Medical College Hospital,Hohhot 010050,China
  • Received:2008-04-18 Revised:2008-08-04 Online:2009-07-11 Published:2009-07-11
  • Contact: DOU Zhi-zhou

摘要: 在细胞彩色图像处理中,为了有效地计算与分析细胞各特征值,对细胞图像的精确的三域分割是细胞自动分析与识别的一个关键环节。提出利用多分类支持向量机对细胞彩色图像进行背景、胞浆与核的一次性三域分割,并且通过聚类分析的方法实现了在线训练,实验表明,该方法在细胞彩色图像的多域分割上,能获得较高的分割精度和较好的鲁棒性。

关键词: 支持向量机, 多分类支持向量机, 一对一, 分割

Abstract: In order to effectively calculate and analyze the character of the cell,it is important to segment accurately cell image into three value map in cell color image processing.This paper proposes using Multi-class Support Vector Machine(M-SVM) to segment the background,nucleus and cytoplasm pixels into three fields in one color image processing.Furthermore,this paper uses clustering method to train M-SVM online.The experiment shows that the method can achieve high accuracy and is robust to varied cell color image acquisition.

Key words: Support Vector Machine(SVM), Multi-class Support Vector Machine(M-SVM), one against one, segmentation