Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (22): 138-139.DOI: 10.3778/j.issn.1002-8331.2009.22.045

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

Ultrasound image de-noising using modified multilevel threshold

NONG Jing-hui1,WANG Jin-li1,LIU Ai-lin2,WANG Xiu-xin1

  1. 1.College of Computer Science and Information Technology,Guangxi Normal University,Guilin,Guangxi 541004,China
    2.Department of Electronic Engineering and Physics,Hunan University of Science and Engineering,Yongzhou,Hunan 425100,China
  • Received:2009-02-02 Revised:2009-04-01 Online:2009-08-01 Published:2009-08-01
  • Contact: NONG Jing-hui



  1. 1.广西师范大学 计算机科学与信息工程学院,广西 桂林 541004
    2.湖南科技学院 电子与物理系,湖南 永州 425100
  • 通讯作者: 农京辉


There are inherent speckle noises in the medical ultrasound image.Image denoising is necessary for lower image quality.Multiwavelet has higher resolution and better denoising effect than scalar wavelet.Therefore,the modified multilevel threshold with fuzzy clustering is proposed based on the energy distribution of multiwavelet coefficients.First the ultrasound image is discomposed to multiwavelet domain.Then its coefficients are separated into noise and signal classes with fuzzy clustering.The signal coefficients are shrunk with different thresholds for image de-noising on the different scales.The result indicates the method is more effective than hard threshold and soft threshold.The speckle noise in the original image is removed efficiently and at the same time the image details are reserved.

Key words: image denoising, fractal interpolation multiwavelet, modified multilevel threshold, fuzzy clustering

摘要: 医学超声图像存在特有的斑点噪声,大大降低了图像质量,必须进行降噪处理。多小波具有比单小波分解更加精确、去噪效果更好的特点。对超声图像进行分形插值多小波分解,根据多小波分解后的能量分布特性,提出了改进多层阈值与模糊聚类相结合方法,将小波系数模糊聚类分成噪声和信号两类,然后在不同尺度对信号小波系数进行不同阈值萎缩处理,实现降噪目的。结果表明该方法优于硬阈值和软阈值法,可有效地降低图像斑点噪声并保留图像细节。

关键词: 图像降噪, 分形插值多小波, 改进多层阈值, 模糊聚类