Computer Engineering and Applications ›› 2013, Vol. 49 ›› Issue (11): 166-168.

Previous Articles     Next Articles

Extraction of texture feature using graph cut method

HUANG Ying1,2,3, LI Weisheng2,3, ZHOU Lifang4, WANG Kuangsheng3   

  1. 1.School of Electronic Engineering, University of Electronic Science and Technology of China, Chengdu 610054, China
    2.Chongqing Key Lab of Computer Network and Communication Technology, Chongqing 400065, China
    3.College of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
    4.College of Software, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
  • Online:2013-06-01 Published:2013-06-14


黄  颖1,2,3,李伟生2,3,周丽芳4,王矿生3   

  1. 1.电子科技大学 电子工程学院,成都 610054
    2.重庆市/信息产业部计算机网络与通信技术重点实验室,重庆 400065
    3.重庆邮电大学 计算机科学与技术学院,重庆 400065
    4.重庆邮电大学 软件学院,重庆 400065

Abstract: Algebraic multi-grid method is analyzed and is applied in the normalized cut method to extract the texture feature of the image. Large grid?density appears in the image regions with radical changes, and small one in the smoother regions. Singularities in the image can be detected by the AMG method, and especially the singularities in the texture image. An energy function is constructed for the texture feature and is minimized using max-flow method. Experimental results show that the proposed method can extract more texture details.

Key words: graph cut method, algebraic multi-grid, max-flow method, texture feature

摘要: 研究目的是对代数多重网格(AMG)方法进行分析,粗网格中会保留强连接部分而去掉弱连接部分,可以提取图像的纹理信息。将AMG方法提取的图像的纹理特征结合到图分割算法中,针对具有纹理特征的图片构建能量函数,并使用最大流方法进行优化。使用一些自然图像进行了验证,结果证明针对该方法能够较好地提取图像的纹理特征。

关键词: 图割方法, 代数多重网格, 最大流方法, 纹理特征