计算机与现代化 ›› 2010, Vol. 1 ›› Issue (01): 113-116.doi: 10.3969/j.issn.1006-2475.2010.01.033

• 图像处理 • 上一篇    下一篇

基于图论的Normalized Cut图像分割方法研究

杨宇鹏,赵卫东,王志成,陈刚   

  1. 同济大学企业数字化技术教育部工程研究中心,上海 200092
  • 收稿日期:2009-06-09 修回日期:1900-01-01 出版日期:2010-01-15 发布日期:2010-01-15

Research on Graph-based Normalized Cut Image Segmentation Method

YANG Yu-peng,ZHAO Wei-dong,WANG Zhi-cheng,CHEN Gang   

  1. Engineering Research Center for Enterprise Digital Technology, Ministry of Education, Tongji University, Shanghai 200092, China
  • Received:2009-06-09 Revised:1900-01-01 Online:2010-01-15 Published:2010-01-15

摘要: 基于图论的图像分割是近年来国际上研究图像分割算法领域中一个新的热点研究方向,Normalized Cut(规一化分割)是基于图论的分割方法中比较典型的一种。本文介绍Normalized Cut方法的相关内容,分析它的优缺点并提出几种有效的改进方法。最后,实验显示Normalized Cut在特定图片的分割上取得了不错的效果。

关键词: 图像分割, 图论, 规一化分割, Nystrom方法

Abstract: Graph-based image segmentation method is newly developing technique in recent years. Normalized cut is one of typically graph-based segmentation methods. This paper introduces correlative content of normalized cut, analyzes advantages and disadvantages of it, and proposes some efficiently solutions for improving. Finally, experimental results show that good result of segmentation can be obtained by the normalized cut algorithm.

Key words: image segmentation, graph theoretic, normalized cut, Nystrom method

中图分类号: