How to generate 3% gaussian noise with N(0,100)

Hi all,

I have seen your explanation about "How to add Gaussian noise to the 1D signal" in Matlab central. It inspired me to ask you about my problem in adding noise in image.

   As I have to compare my result with the result of reference material, I have to use 3% to 8% Gaussian white noise N(0,100)with a synthetic image having only gray values 0 and 90.   I know N(0,100) means the average is 0 and variance is 100.  But I can not understand the 3% noise.

If I use the following command

noiseImage = imnoise (Image,'Gaussian',0,100)

I get an output which is very noisy, and the results as given in the reference material are not reached. I think when imnoise command is used, only the image is transferred to gray scale image ranges between 0 to 1 , and the variance is not scaled down. Any way how to mean the 3% noise is not clear. Herewith I am sending the URL of the reference material for your perusal. I will appreciate if you send your suggestion. Thanks in advance

Reference material: http://parnec.nuaa.edu.cn/zhangdq/pr07.pdf

As I am not to attach and send the pdf of the reference material I am sending the URL here. You can get the reference material on free of cost. In page Number 17 in section 4.1.2 you can see the matter what I want your help