Author Login Editor-in-Chief Peer Review Editor Work Office Work

Computer Engineering

Previous Articles     Next Articles

Image Segmentation Based on Gray-level Spatial Correlation Maximum Between-cluster Variance

HE Jianfeng,FU Zeng,XIANG Yan,YI Sanli,CUI Rui   

  1. (School of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)
  • Received:2014-11-18 Online:2015-11-15 Published:2015-11-13

基于灰度空间相关性最大类间方差的图像分割

贺建峰,符增,相艳,易三莉,崔锐   

  1. (昆明理工大学信息工程与自动化学院,昆明 650500)
  • 作者简介:贺建峰(1965-),男,教授、博士,主研方向:图形图像处理,模式识别;符增,硕士研究生;相艳,讲师、硕士;易三莉,讲师、博士;崔锐,硕士研究生。
  • 基金资助:
    国家自然科学基金资助项目(11265007);教育部留学回国人员科研启动基金资助项目(2010-1561)。

Abstract: When processing an image with the blurred background and target,image segmenting effect by maximum between cluster variance 1D-Otsu and 2D-Otsu is not good.A method for image segmentation based on Gray-level Spatial Correlation(GLSC) maximum between-cluster variance is proposed.The proposed algorithm uses the gray value of the pixels and their similarity with neighboring pixels in gray value to build a histogram which is called GLSC histogram.Then threshold value is obtained by calculating GLSC histogram maximum between-class variance.Integrogram is introduced in order to make the time complexity reduced from original O((N2×L)2) to O(N2×L).Comparing the proposed algorithm with 1D-Otsu,2D-Otsu and GLSC entropy algorithm for segmenting 5 different real images,the proposed algorithm shows better robustness.

Key words: maximum between-cluster variance, Gray-level Spatial Correlation(GLSC), histogram, integrogram, image segmentation, entropy algorithm

摘要: 一维最大类间方差1D-Otsu和二维最大类间方差2D-Otsu在目标和背景比较模糊时,图像分割效果较差。针对该问题,提出一种基于灰度空间相关性(GLSC)最大类间方差的图像分割算法。该算法使用各像素的灰度值与其邻域内相似像素的数目构建直方图,通过计算GLSC直方图的最大类间方差得到分割阈值,应用积分图的思想将运算复杂度由O((N2×L)2)降到O(N2×L),节省了运算时间。针对5幅大小不同和直方图类型不同的真实图像,与1D-Otsu、2D-Otsu和灰度空间相关性熵算法进行分割实验比较,结果表明该算法具有较好的鲁棒性。

关键词: 最大类间方差, 灰度空间相关性, 直方图, 积分图, 图像分割, 熵算法

CLC Number: