计算机应用 ›› 2013, Vol. 33 ›› Issue (11): 3201-3203.

• 多媒体处理技术 • 上一篇    下一篇

基于Mallat-Zhong离散小波变换小波的超声图像各向异性扩散抑噪方法

吴世彬,陈波,董万利,高小明   

  1. 西南科技大学 计算机科学与技术学院,四川 绵阳 621010
  • 收稿日期:2013-05-30 修回日期:2013-07-22 出版日期:2013-11-01 发布日期:2013-12-04
  • 通讯作者: 吴世彬
  • 作者简介:吴世彬(1987-),男,四川广汉人,硕士研究生,CCF会员,主要研究方向:医学图像处理;陈波(1963-),男,四川广汉人,教授,博士,主要研究方向:医学图像处理、嵌入式技术;董万利(1981-),男,四川绵阳人,讲师,硕士,主要研究方向:医学图像处理;高小明(1980-),男,四川广安人,讲师,硕士,主要研究方向:医学图像处理、嵌入式系统技术。
  • 基金资助:
    国家自然科学基金资助项目;四川省科技厅资助项目

Ultrasound image anisotropic diffusion de-speckling method based on Mallat-Zhong discrete wavelet transform wavelet

WU Shibin,CHEN Bo,DONG Wanli,GAO Xiaoming   

  1. School of Computer Science and Technology, Southwest University of Science and Technology, Mianyang Sichuan 621010, China
  • Received:2013-05-30 Revised:2013-07-22 Online:2013-12-04 Published:2013-11-01
  • Contact: WU Shibin

摘要: 针对传统各向异性扩散方法在超声图像散斑噪声抑制中存在的噪声抑制不充分与边缘特征保持不足的问题,提出一种基于Mallat-Zhong离散小波变换(MZ-DWT)小波的散斑噪声抑制方法。该方法将MZ-DWT小波分析与期望值最大化(EM)算法作为图像中均匀区域与边缘区域的鉴别因子,使扩散系数能够更准确地控制扩散强度与扩散速度,从而达到充分抑制噪声和保护边缘的目的。实验结果表明,所提方法在有效抑制散斑噪声的同时,更好地保持了图像细节信息,其性能优于传统各向异性扩散方法。

关键词: 散斑噪声, 各向异性扩散, Mallat-Zhong离散小波变换小波, 期望值最大化算法

Abstract: In view of speckle noise in ultrasound image, there are some disadvantages of traditional anisotropic diffusion methods, such as in-sufficient noise suppression and edge details preservation. A de-speckling method based on Mallat-Zhong Discrete Wavelet Transform (MZ-DWT) wavelet was proposed. The method used MZ-DWT wavelet and Expectation Maximization (EM) algorithm as the discrimination factor between homogeneous and edge regions, making it more accurately to control diffusion intensity and rate and achieving the noise suppression and details preservation. The experimental results show that, the proposed algorithm can better de-speckle while preserving image details and the performance of the method is better than the traditional anisotropic diffusion methods.

Key words: speckle noise, anisotropic diffusion, Mallat-Zhong Discrete Wavelet Transform (MZ-DWT) wavelet, Expectation Maximization (EM) algorithm

中图分类号: