基于小波的医学超声图像斑点噪声抑制方法

斑点噪声是超声图像中固有的噪声。本文提出了一种新的去除斑点噪声的方法，这种方法结合中值滤波和多尺度非线性小波软阈值的优点，首先把原网像进行对数转换，然后把对数转换后的图像进行中值滤波处理，从而把转换后的图像分成两部分，对每一部分进行小波分析，假设小波系数服从广义高斯分布（GGD），利用小波系数的统计特性估计出各个部分各个尺度的阈值，最后用软阈值方法对上述两部分分别去噪。实验结果表明，本文提出的方法在有效去除斑点噪声方面，优于中值滤波，维纳滤波和多尺度非线性阈值算法（MSSNT-A）。

Wavelet-based Method for Speckle Reduction in Medical Ultrasound Image

Speckle is an inherent noise in ultrasound image.A novel speckle-reduction method is presented.This method combined the advantages of median filter and soft threshold process of wavelet sub-band coefficients.At first,the original ultrasound image was transformed logarithmically,then the logarithmically transformed medical ultrasound image was processed into two parts by median filter.Each part was wavelet transformed.The distribution of sub-band coefficients is generalized Gaussian distribution(GGD).The threshold of each scale of each part was estimated by the statistical approach of the wavelet coefficients.At the end,soft threshold method was used to each part for speckle reduction.The experiment results showed the method proposed in this paper was better than median filter,wiener filter and MSSNT-A.