When adding additive white Gaussian noise in MATLAB, one can use the predefined function
J = imnoise(I,'gaussian',M,V) % I is the image to add the noise
with default, zero mean (M) and variance (V) 0.01. The manual for this function is here.
However, in various MATLAB codes, I've also seen that additive Gaussian noise is added to the image by the following way
sigma = 10; % standard deviation (STD)
g = I + sigma * randn(size(I)); %add gaussian noise with STD 10
Which is fine. Now, we know the formula for variance,
[![variance=sigma^2][2]][2]
where sigma is the STD. So, according to the second code, I have sigma = 10 thus, the variance (V) should be 100. Using MATLAB imnoise function for zero mean and variance 100 should be something like this
J = imnoise(I,'gaussian',0,100)
However, this does not produce a corrupted image even close to the second code. The image seems to be 100% corrupted with noise. How is this different ? am I missing something here ?
