Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (35): 182-184.DOI: 10.3778/j.issn.1002-8331.2008.35.054

• 图形、图像、模式识别 • Previous Articles     Next Articles

Efficient method for segmentation of chromosome images

JING Xue,TENG Qi-zhi,TANG Tang,WANG Shu   

  1. Institute of Image Information,College of Electronics and Information,Sichan University,Chengdu 610064,China
  • Received:2007-12-21 Revised:2008-03-25 Online:2008-12-11 Published:2008-12-11
  • Contact: JING Xue

有效的染色体分割新方法

敬 雪,滕奇志,唐 棠,王 姝   

  1. 四川大学 电子信息学院 图像信息研究所,成都 610064
  • 通讯作者: 敬 雪

Abstract: Getting the correct segmentation-point is a major step toward the segmentation of touching and overlapping chromosomes,when analyzing and recognizing chromosome images.This paper applies a calculating method of edge chain-code to find the concave points correctly which segmentation points belong to,that is to say the candidate segmentation points.And then the right ones are estimated and selected by the distance threshold among segmentation points,as well as the arc threshold of edge.At the same time,it introduces a proper theory for two chromosomes which touching head-to-head or head-to-end.

Key words: chromosome, image segmentation, chain code, concave points

摘要: 在染色体图像分析与识别中,将粘连或是交叠的染色体分割开的关键技术是找到正确的分割点。通过使用一种边界链码的计算方法来准确定位分割点所属的凹点,即候选分割点;再利用候选分割点间的距离阈值和边界弧长阈值判断并筛选出正确的分割点。同时提出了两条粘连的染色体在其端部粘连或首尾粘连情况下的正确分割方法。

关键词: 染色体, 图像分割, 链码, 凹点