Gaussian noise: Image Processing
-
28-09-2020 - |
Question
As we all know, Gaussian Noise follows Gaussian or Normal distribution, and that distribution follows a $BELL$ $CURVE$.
As we can see that most of the values are centered around the mean.
Now consider this image
When I add Gaussian noise to this image I get something like this
As we can see that the noise appears to be $UNIFORMLY$ $DISTRIBUTED$ throughout the image. There is no region where we can say that the noises concentrated around the mean value
So how can these be called Gaussian Noise?
The Code that I have used in octave is given below
pkg load image;
I=imread('C:\Users\Hirak\Desktop\apple.jpg');
I=rgb2gray(I);
J = imnoise(I,'gaussian',0.02);
K = medfilt2(J);
imshow(J);
Solution
The noise is Gaussian noise because the values you add to your existing images follow a Gaussian distribution, not the locations of where you add the noise - that is uniform (and not random at all - each pixel gets Gaussian noise added to it).