基于树状小波的超声医学图像阈值函数降噪  

The De-noising of the Ultrasound Medical Image with Threshold Functions Based on Tree-structure Wavelet Analysis

作　　者：王修信[1,2] 胡维平[3] 梁冬冬[3] 覃以威[3] 周建玲[3] 王强[3] 

机构地区：[1]广西师范大学物理与信息工程学院 [2]北京师范大学地理学与遥感科学学院,北京100875 [3]北京师范大学地理学与遥感科学学院

出　　处：《中国医学物理学杂志》2005年第2期479-480,459,共3页Chinese Journal of Medical Physics

基　　金：广西自然科学基金资助项目(桂科自0339036);广西教育厅科研基金资助项目(20022003)

摘　　要：针对超声医学图像中存在特有的斑点噪声,利用树状小波分解比传统小波分解精度高的特点,将超声医学图像进行树状小波分解,然后分别采用硬阈值、软阈值和半软阈值函数三种方法进行降噪处理。结果表明半软阈值函数方法是较优阈值函数方法,可以有效地降低原图像的斑点噪声并保留图像细节。There were inherent speckle noises in the ultrasound medical image. Th e tree-structure wavelet had higher resolution than the conventional wavelet. Th erefore, the ultrasound medical image was discomposed to the tree-structure wave let domain. Then its coefficients were shrunk with the three methods of hard thr eshold, soft threshold, and semi-soft threshold function differently for the ima ge de-noising. The result indicated that the semi-soft threshold was the effecti ve function method. The speckle noise in the original image was removed efficien tly and at the same time the image details were reserved.

关 键 词：超声医学图像 树状小波分析 阈值函数降噪 

分 类 号：R318[医药卫生—生物医学工程]