need help ...sir very urgent pls help me ..an image filter for removing impulse noise removal based on type 2 fuzzy set.need coding

        if true
          % code
        endModification of the current pixel based on type-II fuzzy switching.

Let µ[r(i, j)] Є [0, 1] be the type-I membership function of r(i, j) which indicates the extent of the impulsiveness of the pixel I(i , j). The following fuzzy rules[15] are applied : [Rule 1] If r(i, j) is large, then µ[r(i, j)] is large. [Rule 2] If r(i, j) is small, then µ[r(i, j)] is small. With these rules , the S-function is used to describe the type-I membership function of the impulse noise corruption extent of the current pixel.

(4)

where β = (α + γ) / 2.

268

Figure 4. Upper and Lower limits of a membership function and its type-II construction.

As mentioned before a type-II fuzzy set may be obtained by blurring a type-I membership function. For this purpose we use our type-I fuzzy set and assign upper and lower membership degrees to each element to construct the FOU as shown in figure 4. A more practical definition for a type-II fuzzy set can be given as follows[22]-

(5) The upper and lower membership degrees of the initial membership function µ can be defined as -
µupper(x) = (µ(x))0.5  (6)
µlower(x) = (µ(x))2  (7)
where 0 ≤ µ(x) ≤ 1 is the type-I membership function for value x (Figure 4).
After calculating µ(x) using equation (4), we calculate
values of µupper(x) and µlower(x) from equations (6) and
(7) respectively. Now the FOU is bounded by upper and lower membership functions. Here the uncertainties in the shape and position of the type-I fuzzy set can be represented by means of the third dimension represented by the FOU.
The proposed type-II membership function is:  (8)  
µT2[r(i, j)] = (µupper x δ) + (µlower x (1 – δ))    
where 0 ≤ δ ≤ 1  and    
δ = (│gmean – (L/2)│) / (L/2)  (9)  
where gmean is the mean value of each sub-image and L is the number of gray levels used. Finally, using µT2 the
proposed type-II filter calculates the output value as follows:
Y(i, j)= (µT2 X Med ) + ( I(i, j) X (1 - µT2 ))  (10)
The idea behind the µT2 equation is that pixels with extreme values (pixel values in a darker or a lighter neighbourhood) will have greater proportion of µupper

than µlower.