提出一种有效的语音增强算法，该算法使用重叠相加FF-r滤波器进行谱幅度估计，给出噪声估计的两个均值参数，此参数适用于任何噪声．而且进行噪声估计无需检测语音间隙，无需VAD的运算过程，克服了单信道噪声 压缩算法的缺点．-An effective speech enhancement algorithm using the overlap-add FF-r filter in spectral amplitude estimation, given the noise mean estimate

This code is for modified mean filter,here the noise can be removed,this is FOR RGB images