Generate the optimum coefficient using optimization filter

Dear all

I have a gray image . I convert it to DCT . Then i take 4x4 blocks from this DCT of the gray image . I convert the 4x4 dct blocks to idct . Then i filter this 4x4 using optimization filter . . I chose the coefficient of the filter to be a(1)=1/7; a(2)=1/7; a(3)=1/7; a(4)=1/7; a(5)=1;

However, my supervisor want the optimization filter to generate the optimum coefficient itself .So how i could make the code to chose the optimization filter ?. Is this optimization filter is correct ? I used to m-file one for Global and one for the optimization filter The code for optimization part is as the following :

 % clear all
close all
clc
global image1 image2

fname1=input('Enter image file name ? ','s')

[x1]=imread(fname1);

 k=input('Enter k? ');

len1=length(x1);

 imm = x1;

 im = ones(size(imm));
 im_1 = uint8(im);
 s= 1;
 for s=1:3
green1=x1(:,:,s);

gray1=rgb2gray(x1);

% doubel gray2=double(gray1); green2=double(green1);

% Convert to DCT gray3=dct2(gray2);

green3=dct2(green2);

%DCT_green=green3(1:4,1:4);

tred=ones(len1,len1);

    for i=1:k
        for j=1:k

freqblockred(i,j)=green3(i,j);

freqblockgray(i,j)=gray3(i,j);

        end
    end

 IDCT_Gray=idct2(freqblockgray);
 IDCT_green=idct2(freqblockred);
 green5=double(IDCT_green);

% The optimization filter coefficients: 
a(1)=1/7;
a(2)=1/7;
a(3)=1/7;
a(4)=1/7;
a(5)=1;

image1=IDCT_Gray;

image2=IDCT_green;

[zy,mse]= fminsearch(@routine, a);

h(1)=zy(1);

h(2)=zy(2);

h(3)=zy(3);

h(4)=zy(4);

h(5)=zy(3);

h(6)=zy(2);

h(7)=zy(1);

hh=transpose(h)*h;

yy2=zy(5)+imfilter(IDCT_Gray,hh,'replicate');

yy3=double(yy2); yy4=dct2(yy3);

gray4=gray3; gray4(1:4,1:4) = yy4;

gray5 = idct2(gray4); gray6 = double(gray5); %Calculate MSE between the original red layer & generated one [row1,col1] = size(gray6); diff1=(uint8(gray6)-uint8(green2)).^2; mes1=sum(sum(diff1))/(row1*col1); psnr =10*log10((255*255)/mes1)

imm9 = uint8(gray6);

 im_1(:,:,s) = imm9 ;
 end
psnr2 = 10*log10((255*255)/((sum(sum((x1-im_1).^2)))));
imshow(uint8(im_1));

while the code for global is as the following :

function mse=routnie(a)

global image1 image2

%--- Filter

% 1D filter

b(1)=a(1); b(2)=a(2); b(3)=a(3); b(4)=a(4); b(5)=a(3); b(6)=a(2); b(7)=a(1);

%2D Filter

hh=transpose(b)*b;

image3=a(5)+imfilter(image1,hh,'replicate');

image4=uint8(image3);

image5=double(image4);

%-------------

mse=((sum(sum((image5-image2).^2))));

return

So please let me know how to solve the problem of finding the optimum coefficients that will give me maximum PSNR .. thanks

[zgrt]

please any one can answer my question .. I want to find the optimum coefficients that will give maximum PSNR between gray image & the filtered image .