Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (35): 194-196.DOI: 10.3778/j.issn.1002-8331.2008.35.058

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

Three numerical value schemes of curves evolving based on mean curvature motion

WANG Lei,ZHU Xuan,ZHANG Shen-hua   

  1. College of Information Science & Technology,Northwest University,Xi’an 710127,China
  • Received:2007-12-19 Revised:2008-03-03 Online:2008-12-11 Published:2008-12-11
  • Contact: WANG Lei

曲率驱动的曲线演化的三类数值方案

王 蕾,祝 轩,张申华   

  1. 西北大学 信息科学与技术学院,西安 710127
  • 通讯作者: 王 蕾

Abstract: Three numerical calculating schemes of realizing the curvature motion formula in curve evolving are discussed,including explicit scheme,semi-implicit scheme and median filter scheme.Furthermore compares their evolvement goal,evolvement precision and time expense through the experiment research.And the result shows that the relaxation algorithm precedes the other algorithm whatever in evolvement goal,precision and speed.

Key words: mean curvature motion, curves evolving, explicit scheme, semi-implicit scheme, discrete median filter

摘要: 探讨了三类实现曲线演化的曲率运动方程的数值计算方案:显示方案、半隐式方案和中值滤波方法。进一步通过实验研究并比较了它们的演化目标、演化精度和时间代价等方面的性能。结果表明半隐式方案中的超松弛算法无论是从演化目标、演化精度还是演化速度等方面均优于其它方案。

关键词: 平均曲率运动, 曲线演化, 显式方案, 半隐式方案, 离散中值滤波