基于平衡正交多小波的MR图像去噪算法研究  

MR Image De-nosing Based on Balanced Orthogonal Multiwavelet

作　　者：张艳华[1] 李玲远[2] 

机构地区：[1]徐州工程学院信电工程学院,江苏徐州221008 [2]华中师范大学电子信息与工程系,湖北武汉430079

出　　处：《通信技术》2009年第3期236-238,共3页Communications Technology

摘　　要：磁共振图像中的噪声分布的特殊性不仅大大降低了图像的质量,影响医学诊断,而且为噪声的去除提出了更高的要求。平衡正交多小波同时满足正交性和对称性,不仅具有比单小波分解更加精确、去噪效果更好的特点,同时避免了普通多小波分解过程中的预处理过程。文中对磁共振图像进行平衡正交多小波分解后,采用了一种自适应性更高的阈值确定函数与半软阈值相结合方法对小波系数进行萎缩处理,实现降噪目的。结果表明:该方法优于97提升小波及ghm多小波方法,去噪效果进一步提高。The particular noise in MRI not only lowers the quality of image but also makes the denoising more difficult. The balanced orthogonal multiwavelet with the orthogonality and symmetry decomposes the image more precisely without pre processing and denoises better than single wavelet. Based on decomposition of the image by balanced orthogonal multiwavelet, a method in combination of threshold-selecting function with semi-soft thresholding is proposed to reduce the noise in MRI. The simulation results shows that the method is superior to others.

关 键 词：磁共振图像 去噪 平衡正交多小波 自适应阈值 半软阈值函数 

分 类 号：TN911.7[电子电信—通信与信息系统]