高斯核函数对高斯白噪声进行卷积滤波
 

各位大家好！现在有一个问题急需解决，请大家帮忙！原文是这样说的：1000 two-dimensional SPIs were generated using Matlabs random Gaussian field generator.The SPIs have zero mean and unit standard deviation. Each SPI consists of either 128 x 128 pixels. The SPIs were filtered using a Gaussian kernel with an FWHM equal three pixels 请问这个该怎么实现啊？我的code是这样写的:
a=randn(128,128,1000);
fwhm=3;
s=fwhm/(2(sqrt(2*log(2))));
w=fspecial('gaussian',5*fwhm,s);
b=imfilter(a,w,'conv');
得出的结果是b的最大值几乎都在[-1,1]之间，可是按要求应该有大于3的值，请大家帮帮忙？看看哪里出错了！万分感谢！