What is the mathematics behind the imnoise function in matlab?

Actually I want to know that how imnoise add noise to a signal .which method it follows.

Answer 1

[Walter Roberson]

It depends which kind of noise. http://www.mathworks.com/help/images/ref/imnoise.html

Most of the supported types are essentially randn() * scale + mean

[Ashwani]

Thanks for reply. If I am using J = imnoise(I,'gaussian',m,v),( m = 0,v=1). Then how it add noise to 'I'.

[Jurgen]

I would guess newpixel = originalpixel + randn(1). For every pixel in image, in other words J = I + randn(size(I)). See 'help randn' on how to get different means/variance.

Answer 2

[Image Analyst]

This is all spelled out in the help for the different types of noise you can choose from. It assumes the image is normalized 0-1 and then adds the noise you selected, then it unnormalizes the image to get back to the original intensity range. Was there one particular type of noise that you did not understand the definition of?