difference between two programs

I have two small programs which must have the same result but in executing them I have some difference and I don't understand where is the problem. The task is to write a simple stand-alone program that converts the loaded source image in the negative one.

X=imread('mosque.jpg');
imshow(X);

[n,m]=size(X);
B=zeros(n, m, 'uint8');
for i=1:n
    for j=1:m
        B(i,j)=255-X(i,j);
    end
end
subplot(2,2,1);
imshow(X);
subplot(2,2,2);
imhist(X);
subplot(2,2,3);
imshow(B);
subplot(2,2,4);
imhist(B);
    

the second program without the for

X=imread('mosque.jpg');
imshow(X);

[n,m]=size(X);
B=zeros(n, m, 'uint8');

        B=255.-X;
   
subplot(2,2,1);
imshow(X);
subplot(2,2,2);
imhist(X);
subplot(2,2,3);
imshow(B);
subplot(2,2,4);
imhist(B);

[John D'Errico]

Note that for these two statements:

B=zeros(n, m, 'uint8');
B=255.-X;

The first is irrelevant. There is no need to preallocate B there, because the second line just overwrites it.

Had you done this instead?

B = 'Mississippi';
B=255.-X;

The result would have been the same.

However, in the case of the for loops, since there you are populating B one element at a time, it is very important to preallocate B.

[boutheina gherib]

when I put this instruction I preallocate B or not

B=zeros(n, m, 'uint8');

in addition I dont understand what do you mean by

B = 'Mississippi'

The result of the version using for is the same photo in gray but in three example and without for the result is one image in gray

Answer 1

[Walter Roberson]

X=imread('mosque.jpg');

The great majority of .jpg are RGB (grayscale is possible but very rare.)

[n,m]=size(X);

size() of an RGB image when permitting only two output variables would store the number of rows in n, and three times the number of columns in m . When you use size() with fewer outputs than dimensions of the array, the result is not to throw away the extra dimensions. The product of the values returned by size() is always equal to the number of elements in the entire array.

B=zeros(n, m, 'uint8');

B is constructed as a grayscale image, not as an RGB image.

[boutheina gherib]

I don't understand exactly what do you mean I must use size without argument or what

[Walter Roberson]

>> A = randi(10,5,6,7);
>> B = size(A)

B =

     5     6     7

>> [C,D] = size(A)

C =

     5


D =

    42

>> [E,F,G] = size(A)

E =

     5


F =

     6


G =

     7

Notice that the [C,D] = size(A) case did not return C = 5 and D = 6. These two pieces of code are not equivalent:

     [C,D] = size(A);      %D will be 6*7 = 42. C*D will equal numel(A)
     %versus
     temp = size(A);
     C = temp(1); D = temp(2);  %D will be 6. C*D will equal 30 only

You can pre-allocate B as:

     B = zeros(size(A), class(A));

But remember in your loops that A is not two dimensional: A is three dimensional. You have to loop over the red, green, and blue color planes.

Answer 2

[boutheina gherib]

in conclusion I can't modify this because from the start the type of the image jpg give three time the output

Answer 3

[boutheina gherib]

Thank you for your answer I make some change in my two programs

and the problem is solved

X=imread('mosque.jpg');
imshow(X);

[n,m,p]=size(X);
B=zeros(n, m,p, 'uint8');
for i=1:n
    for j=1:m
        for k=1:p
        B(i,j,k)=255-X(i,j,k);
        end
    end
end
subplot(2,2,1);
imshow(X);
subplot(2,2,2);
imhist(X);
subplot(2,2,3);
imshow(B);
subplot(2,2,4);
imhist(B);

without for

X=imread('mosque.jpg');
imshow(X);

[n,m,p]=size(X);
B=zeros(n, m,p, 'uint8');

        B=255.-X;
   
subplot(2,2,1);
imshow(X);
subplot(2,2,2);
imhist(X);
subplot(2,2,3);
imshow(B);
subplot(2,2,4);
imhist(B);