基于小波分析的磁共振图像去噪方法研究  

作　　者：董凯宁[1] 张松[2] 

机构地区：[1]四川大学电子信息学院 [2]成都电子机械高等专科学校通信工程系

出　　处：《成都电子机械高等专科学校学报》2007年第3期1-5,共5页Journal of Chengdu Electromechanical College

摘　　要：小波分析去噪是一种新兴的图像去噪方法。由于小波分析具有多尺度分辨和去相关性等特点,使得小波分析在去除高斯白噪声方面优于传统的图像去噪方法。但是,磁共振图像这种小波变换去噪后,纹理特征被弱化,图像的边缘变得模糊。针对以上问题,提出了基于小波变换的图像去噪新方法。此去噪方法对经典的小波去噪方法进行了改进,使基于阀值的小波分析在阀值处理上更精确,并具有自适应性。采用本方法处理的噪声图像与经典方法相比,图像的边缘信息更清晰,纹理特征增强,去噪能力也得到增强。Image denoising with wavelet analysis has been flourishing. With thecharacters of multi - scales and de - relevence, wavelet analysis is better than typical methods in magnetic resonance image denoising of Gauss noise. But, with wavelet analysis, texture characters of magnetic resonance images are weakening, edges of images are blurred. To solve these problems, a new magnetic resonance image denoising method based on wavelet analysis is proposed. This method is a improvement one based on typical wavelet analysis, and making the threshold more accuracy than the one in typical method and has self - compatibility. Compared with classical wavelet denoising method, this method makes edges more clear, enhancestexture characters and improves the ability of denoising.

关 键 词：磁共振图像 小波分析 去噪 

分 类 号：TN911.73[电子电信—通信与信息系统]