Question

请求高手翻译成英文，非常感谢！

极值中值滤波（EM）用于滤除椒盐噪声时，首先根据判别标准（是否为极值点）将全部象素分为两类：噪声N与信号S。然后，根据空间相关性将噪声咪由该点领域的中值取代。EM算法的执行速度与标准中值滤波相当，但保留了较多的细节信息，没有造成图象的模糊，其滤波效果在输入图象信噪比大于0时与PEM相当，但不需迭代，执行速度提高了若干倍。在输入信噪比更低的情况下，只需迭代一次，就能取得与PEM相当的输出效果。我们根据经验可以知道，对于一幅自然图像，相邻点之间存在着很大的相关性。图像中某一点的灰度值与其周围点的灰度值非常接近。除了孤立点一般认为是噪声外，即使在边缘部分也是这样。根据这一点可以判断，在一幅图像中如果一个像素点的值远大于或小于其邻域的值，也就是说，该像素与其邻域的相关性很小。那么，该点很可能己被噪声污染了。否则，如果其值与其邻点值接近，就应该是一个有效的信号点。SM方法门限需要事先确定，而门限值的选择又与所处理图像的噪声密度有关，由于对不同的区域采用相同的门限也不甚合理，所以结果并不令人十分满意，尤其是在噪声密度较大的情况。

Answer 1

Extreme median filtering (EM) used to filter out the salt &pepper noise, the first according to the judgment standard (whether for extreme value point) will all pixels are divided into two types: noise N and signal S. Then, depending on the spatial correlation noise by that point in the field of mi value instead. EM algorithm, speed and the execution of the standard median filtering quite, but retained more detail information and causes no images, the effect of the fuzzy filter in the input image signal-to-noise ratio greater than 0, but with a PEM don't need iteration, execution speed increase a number of times. In the input SNR lower case, just iteration time, can achieve the output effect and a PEM. We can know, according to experience for a natural images, between adjacent point there is a great correlation. The image of a little gray value and