the problem with loop

test1 images and reference images imagedb exceeds 6 mb size so i couldn't attach the image for test you can take any images for test

the problem is am getting only surface roughness 1 is displayed

am unable to get the rest surface roughness

test1=imread('D:\HD\Images\5 0.75 30\5 0.75 30 4.bmp');

imagedb{1}=imread('D:\HD\Images\5 0.75 30\5 0.75 30 1.bmp');
imagedb{2}=imread('D:\HD\Images\5 1.25 30\5 1.25 30 1.bmp');
imagedb{3}=imread('D:\HD\Images\10 0.25 30\10 0.25 30 1.bmp');
imagedb{4}=imread('D:\HD\Images\10 0.75 30\10 0.75 30 1.bmp');
imagedb{5}=imread('D:\HD\Images\10 0.75 50\10 0.75 50 1.bmp');
imagedb{6}=imread('D:\HD\Images\15 1.25 30\15 1.25 30 1.bmp');

resizetest=imresize(test1,[256,256]);
eqtest=histeq(resizetest);
gtest=rgb2gray(eqtest);

imhist(eqtest)
%imhist()

for x=1:6

      resize{x}=imresize(imagedb{x},[256,256]);

eq{x}=histeq(resize{x});

g{x}=rgb2gray(eq{x});

end
%count=0;
%rg{1}=imresize(g{1},[20 20]);
%rgtest=imresize(gtest,[20 20]);
%hamming=pdist2(rg{1},rgtest,'hamming');
%disp(hamming)
%for i=1:20
 %  for j=1:20
%count=hamming(j,i)+count;
 %  end
%end
%answer=count/400;

%disp(answer)
min=10^5;
for i=1:6
    dist{i}= (1/6) * (sum((g{i} - gtest)) );
    if dist{i}< min
    min=dist{i};
    index=i;
    end
end

disp(min)

disp(index)

switch(index)
    case 1
        fprintf('The Surface Roughness Is 2.403\n');
        case 2
            fprintf('The Surface Roughness Is 3.22\n');
            case 3
                fprintf('The Surface Roughness Is 1.595\n');

                  case 4 
                      fprintf('The Surface Roughness Is 2.236\n');

                      case 5
                          fprintf('The Surface Roughness Is 2.125\n');
                          case 6
                              fprintf('The Surface Roughness Is 1.055\n');
  end

%Dist1 = (1/6)*(sum((g1(:).* g2(:) );
%Dist2 = (1/6)*(sum((g1(:).* g3(:) );
%Dist3 = (1/6)*(sum((g1(:).* g2(:) );

%disp(Dist1)
%disp(Dist2)
%disp(Dist3)

[Walter Roberson]

histeq() with one parameter is only valid for intensity (grayscale) images, and in that case returns the equalized grayscale image. You record that in eq{x}. But why do you rgb2gray() that, since you can be sure it is already gray?

[juveria fatima]

yeah ok that i will change ,please help me with the loop

Answer 1

[Walter Roberson]

You have

 dist{i}= (1/6) * (sum((g{i} - gtest)) );

where g{i} is a grayscale version of an image that was resized to 256 x 256 and then histogram equalized. Therefore g{i} will be a 2D array, 256 x 256. When you sum g{i}-gtest you are summing over the first dimension and will get a 1 x 256 result. dist{i} is therefore 1 x 256.

You then have

    if dist{i}< min
    min=dist{i};
    index=i;
    end

but remember dist{i} is a 1 x 256. When you < a 1 x 256 against a scalar, you get a 1 x 256 result. When you if a vector, the result is considered true only if all entries are non-zero, so your test is equivalent to if all(dist{i} < min) . And perhaps that only ever happens for i == 6 -- the other i values might lead to vectors in which some of the entries are less than the minimum but not all of them are.

Your code does not promise that index will be set at all. You only set index conditionally, and it could be the case that none of the conditions match.