Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (19): 196-199.

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

Video object segmentation based on MRF model using wavelet transform

XIE Lei,LI Mei,GAO Zhi-yong,LIU Hai-hua   

  1. College of Electronic and Information Engineering,South-Central University for Nationalities,Wuhan 430074,China
  • Received:2007-09-27 Revised:2007-12-11 Online:2008-07-01 Published:2008-07-01
  • Contact: XIE Lei

一种基于小波变换的马尔可夫随机场的视频对象分割

谢 磊,李 梅,高智勇,刘海华   

  1. 中南民族大学 电子信息工程学院,武汉 430074
  • 通讯作者: 谢 磊

Abstract: Markov Random Field(MRF) theory has widely been applied to segmentation in video images.In this paper,a video object segmentation algorithm based on Markov Random Field using wavelet transform is proposed.After constructing the image pyramid by wavelet transform,energy function of MRF model is defined.The authors solve the superior solution of energy function by reiteration,get a label field,and then extract the moving object.Experimental results are provided using trevor and miss american image sequences.The results show that the proposed algorithm can restrain noise effectively and extract the moving object exactly.

Key words: Markov Random Field(MRF), wavelet transform, video object segmentation

摘要: 马尔可夫随机场(Markov Random Field,MRF)理论已经被广泛地应用于视频图像的分割。提出一种基于小波变换的马尔可夫随机场模型的视频对象分割算法。该算法利用小波变换将图像序列分解到小波域,并在此基础上建立马尔可夫随机场模型,构造相应的能量函数。通过迭代求解能量函数的最优解,得出标记场,提取出运动对象。仿真结果表明,该算法能够有效地抑制噪声,提高构成对象边界像素的数量,快速有效地提取出视频对象。

关键词: 马尔可夫随机场, 小波变换, 视频图像分割