一种基于非线性域值小波包的超声图像闪斑抑制方法

在B超三维成像中，采用了一种新的基于非线性域值小波包的闪斑（Speckle)噪声抑制方法来提高成像质量。首先将一幅原始图像经过预滤波分解为高频噪声图像和低频主体图像两部分，然后分别经过多分辨率小波包处理，得到一组小波包域的系统。该系统采用一种软阈值滤波算法处理，去除掉其中的噪声并保留下各自的有用信息。两组分别去噪后的小波包系数经反变换回到时域并相加，得到消除了噪声的图像。该算法的性能在计算机仿真图像上进行了验证。采用该方法处理了一组水槽中的碗状物体的B超图像，并进行了三维体视化重构。实验表明，这种方法可以有效地抑制超声闪斑噪声并在重构中保留对象的细节特征。

A SPECKLE SUPPRESSION METHOD BASED ON NONLINEAR THRESHOLD WAVELET PACKET IN ULTRASOUND IMAGES

A novel method based on nonlinear threshold wavelet packet was explored in order to suppress the speckles in B scan 3 D ultrasound imaging. Firstly, an original image was separated into two parts by adaptive pre filtering. These two parts were then transformed into a multiscale wavelet packet (WP) domain and the WP coefficients were processed by a soft thresholding filter method. The processed coefficients for each part were then transformed back into the space domain. Finally, the de noised image was obtained as the sum of the two processed parts. Performance of this new method has been tested by a computer simulated image. Ultrasound images set of a plastic bowl in a water tank has been utilized for 3 D ultrasound reconstruction. It was found that this technique reduces the speckle noise effectively, and preserves the resolvable details.