Let us develop one specific algorithm. We assume that the additive noise is zero mean and white with variance

Its power spectrum is then given by,

Consider a small local region in which the signal is assumed stationary within the local region, the signal is modeled by,

where are Local mean and standard deviation of and is zero-mean white noise with unit variance. The signal in the above expression is actually modeled by sum of a space-variant local mean and white noise with a space variant variance within the local region, then, the Weiner filter is given by,

=

i.e is a scaled impulse response given by .

Using the above equation and Fig (6.3 ), the processed image within the local region can be expressed as,

If we assume and are updated at each pixel,

	(6.3.1)

The algorithm based on (6.3.1) can be viewed as a special case of a two channel process. In the two channel process, the image to be processed is divided into two components, the local mean and the local contrast

The local mean and local contrast are modified separately and then the results are combined. In equation (6.3.1) the local mean remains unaltered while the local contrast is scaled according to the relative amplitudes of and