• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2007, Vol. 29 ›› Issue (5): 66-68.

• 论文 • 上一篇    下一篇

一种基于遗传算法的FFT Snake模型图像分割方法

陈勤 刘茵 王涛   

  • 出版日期:2007-05-01 发布日期:2010-06-02

  • Online:2007-05-01 Published:2010-06-02

摘要:

本文针对Snake模型用于轮廓跟踪时存在抗噪性能差、易于从弱边界溢出的不足,对其能量函数进行改进,提出一种新的FFT Snake模型。该模型较好地解决了以上问题,并将 FFT Snake模型的解作为遗传算法的搜索空间,利用遗传算法的全局优化性能,有效地克服了Snake轮廓局部极小化的缺陷,从而可得到对目标更精确的分割。实验结果表明,该方法分割效果十分理想。

关键词: Smke模型 FFT Snake模型 图像分割 遗传算法

Abstract:

This paper presents a new FFT Snake model by improving the energy of the Snake Model, and solves the problem of Snake Model which has a bad anti-noise ability and tends to overflow from the weak edges when applied to contour tracking, Then it considers the solution of the FFT Snake Model as a searchin g space and avoids local minima by making use of the global optimal ability of the genetic algorithm, and obtains more precise segmentation. The experim ental results indicate that the effect of the segmentation is ideal.

Key words: (Snake model, FFT Snake model, image segmentation, genetic algorithm)