混合高斯噪声下的医学荧光染色图像去噪

中值滤波是比较常见的去噪方法,但如何选择最优滤波窗口则是决定图像去噪效果的关键性因素,尤其是针对无法获知无噪声的真实图像的医学图像,通过估计平滑后的误差来选择滤波窗口变得更加困难。本文通过Pilot估计针对医学荧光染色图像中的混合高斯噪声进行了估计,再通过广义SURE方法计算不同窗口下的估计误差,从而得到中值滤波的最优窗口大小,使得图像在尽可能保留纹理信息的情况下最大可能的去除了图像噪声。

Fluorescence Ratio Medical Image Denosing Under Gaussian Mixtures Noises

Median filtering is a popular tool for image denosing, and the selection of the span is the most important factor to the denosing result, especially with medical images whose real image without noise is difficult to acquire. In this letter, we estimated the parameters of the Gaussian mixtures of the fluorescence？ratio？medical image, then found out the optimal span of the moving window and obtained the best result with sharp edges saved.