基于小波软阈值滤波的医学图像去噪  

Medical Image Denoising Based on Wavelet Soft Threshold Filting

作　　者：傅伟[1] 万洪晓[2] 熊平[3] 

机构地区：[1]江苏财经职业技术学院计算机工程系,江苏淮安223003 [2]临沂市人民医院影像中心,山东临沂276003 [3]中南大学信息物理工程学院,湖南长沙410083

出　　处：《中国医学物理学杂志》2009年第1期982-984,共3页Chinese Journal of Medical Physics

摘　　要：目的:为了更好的去除DR医学图像噪声。方法:通过分析其噪声来源,在小波去噪的基础上进行改进。引入方差不变性变换来调整原始图像的噪声模型为高斯噪声模型。图像分解为不同频率的不同子带的小波系数,分别进行不同阈值的滤波。结果:与普通的全局小波去噪方法相比,该方法不但可以保留图像的边缘信息,而且能提高去噪后图像的峰值信噪比。结论:用此方法处理DR图像在噪声去除、细节质量及骨骼锐化等方面比传统的高斯滤波及小波全局阈值滤波等方法效果要好。Objective： To denoise digital radiographic images well. Methods： The paper presents a technique that uses the Anscombe＇s transformation to adjust the original image to a Gaussian noise model based upon the wavelet denoising method. The image is decomposed in different subbands of frequency and orientation and these coefficient are tilted using different thresholds. Results： The proposed method that could keep images edges from damaging and increase PSNR is better than the whole wavelet denoising method. Conclusions： Quantitative and qualitative DR images assessment showed that the proposed algorithm outperforms the traditional Gaussian filter in terms of noise reduction, quality of details and bone sharpness.

关 键 词：小波变换 方差不变性变换 峰值信噪比 医学图像去噪 

分 类 号：R197-39[医药卫生—卫生事业管理]