计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (31): 191-193.

• 图形、图像、模式识别 • 上一篇    下一篇

小波神经网络在图像配准中的应用

舒小华1,2,沈振康1,曾广胜3,龙永红2   

  1. 1.国防科技大学 ATR实验室,长沙 410073 
    2.湖南工业大学 电气与信息工程学院,湖南 株洲 412008
    3.湖南工业大学 包装新材料与技术重点实验室,湖南 株洲 412007
  • 出版日期:2012-11-01 发布日期:2012-10-30

Wavelet neural network application in image registration

SHU Xiaohua1,2, SHEN Zhenkang1, ZENG Guangsheng3, LONG Yonghong2   

  1. 1.ATR Lab, National University of Defense Technology, Changsha 410073, China
    2.College of Electrical and Information Engineering, Hunan University of Technology, Zhuzhou, Hunan 412008, China
    3.Key Lab of New Material and Technology for Package, Hunan University of Technology, Zhuzhou, Hunan 412007, China
  • Online:2012-11-01 Published:2012-10-30

摘要: 提出了一种基于小波网络的图像配准方法。将特征点定义进行了推广,提出了一种以特征区域定义和提取方法。使用Zernike矩表征区域的特征并进行特征区域的对应。因图像配准变换是复杂且难以预知的,利用小波神经网络具有良好的函数逼近性能,提出了具有局域特性的小波神经网络模型逼近图像的配准变换。实验表明这是一种有效的图像配准方法。

关键词: 图像配准, 特征区域, Zernike矩, 小波网络

Abstract: A method for image registration based on wavelet neural network is proposed. The definition of feature points is extended, and a method of definition and extraction of feature region is put forward. The regional characteristic is charactered and corresponded with Zernike moments. For image registration transformation is complex and unpredictable, wavelet neural network has a good approximation performance in the local domain, so it is applied to simulating image registration transform. The experiments show that it is an effective method for image registration.

Key words: image registration, feature region, Zernike moment, wavelet neural network