计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (26): 186-187.DOI: 10.3778/j.issn.1002-8331.2008.26.057

• 图形、图像、模式识别 • 上一篇    下一篇

基于非抽样Contourlet变换的医学图像融合算法

何国栋,于 梅,殷 兵,梁 栋   

  1. 安徽大学 计算智能与信号处理教育部重点实验室,合肥 230039
  • 收稿日期:2008-03-31 修回日期:2008-06-10 出版日期:2008-09-11 发布日期:2008-09-11
  • 通讯作者: 何国栋

Fusion of medical images based on Nonsubsampled Contourlet Transform

HE Guo-dong,YU Mei,YIN Bing,LIANG Dong   

  1. Educational Department Key Lab. of IC&SP,Anhui University,Hefei 230039,China
  • Received:2008-03-31 Revised:2008-06-10 Online:2008-09-11 Published:2008-09-11
  • Contact: HE Guo-dong

摘要: 分析了非抽样Contourlet变换(Nonsubsampled Contourlet Transform,NSCT)的原理,提出了一种新的基于NSCT的医学图像融合算法,应用NSCT对CT和MRI图像进行多尺度、多方向分解,低频子带采取区域能量加权法融合,带通子带采取模最大融合,最后将融合的系数进行NSCT逆变换得到融合图像。实验表明,与其它融合算法比较,该算法融合图像效果较好。

Abstract: The principle of Nonsubsampled Contourlet Transform(NSCT) is studied.A novel fusion algorithm is proposed based on NSCT.Firstly,the CT and MRI images are decomposed at different scales and directions,the low frequency subband coefficients and the the bandpass subband coefficients are obtained.Secondly,different fusion rules are applied in different subband coefficients.The regional energy fusion rule is applied in low frequency subband coefficients and the maximum module rule is applied in the bandpass subband coefficients.Finally,the fused image is reconstructed by the inverse NSCT.The experimental results show that the qualitative of the novel algorithm is better than other fusion algorithms.