基于自适应邻域的超声图像加权中值滤波

A Weighted Median Filter for Medical Ultrasound Image Based on Adaptive Neighborhood

提出了一种自适应邻域中值滤波算法 ,用于医学超声内窥镜图像的噪声滤除。该方法以图像象素邻域的灰度方差为阈值 ,进行保持与修复窗口的自适应改变 ,在有效抑制Speckle噪声的同时 ,较好保留了图像的细节信息。对本算法与Loupas提出的加权中值滤波算法进行了比较 ,指出本算法在一定程度上克服了加权中值滤波器的不足并保留了它的优点 ,对超声内窥镜图像的滤噪有较好的效果。

This paper presented a new weighted median filter for medical ultrasound endoscopes images based on adaptive neighborhood. The method uses histogram square error to choose adaptive median filter, so as to remove most of the noise and keep details of the images. The property of the new filter was analyzed and the medical ultrasound endoscopic filter image with good quality was obtained.