计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (32): 182-184.DOI: 10.3778/j.issn.1002-8331.2010.32.051

• 图形、图像、模式识别 • 上一篇    下一篇

C-V活动轮廓模型的一个注记

申小娜,林 琼   

  1. 后勤工程学院 基础部,重庆 400016
  • 收稿日期:2009-03-26 修回日期:2009-06-01 出版日期:2010-11-11 发布日期:2010-11-11
  • 通讯作者: 申小娜

Remark on C-V active contours model

SHEN Xiao-na,LIN Qiong   

  1. Department of Fundamental Studies,Logistical Engineering University,Chongqing 400016,China
  • Received:2009-03-26 Revised:2009-06-01 Online:2010-11-11 Published:2010-11-11
  • Contact: SHEN Xiao-na

摘要: Chan-Vese提出的“无边活动轮廓”模型(C-V模型)是一个著名的基于区域的图像分割模型,它是基于Mumford-Shah泛函和二值PC函数(目标区域取一个值,背景区域取另一个值)解决图像分割问题的。在C-V模型中,定义能量泛函的面积项的系数被要求为非负值,这个要求限制了模型适用的范围。实验研究表明:面积项系数取负值时,C-V模型能够分割某些原来不适用的图像。

Abstract: The “active contours without edges” model proposed by Chan and Vese(C-V model) is one of the well-known region-
based image segmentation models.The C-V model employs the well-known Mumford-Shah functional to deal with the problem of image segmentation by using piecewise constant representation.In this model,the weight parameter of the area term in the energy functional is chosen to be a nonnegative constant;however,this restriction may limit the applicability of the model in some applications.The empirical research shows that the C-V model with negative area parameter can segment some images that the original C-V model is not applicable to.

中图分类号: