How can I vectorize the following code?

main.m

I = imread('bear.png');
J = rgb2gray(imread('marked_bear.png'));

mask = I-J;

for r=1:size(I,1)
    for c=1:size(I,2)
         if(mask(r,c)) 
             I = some_function(I, r, c);
         end
    end
end

imshow(I);

some_function.m

function I = some_function(I, r, c)
      % some processing on image 'I'
      I(r,c) = 255;

bear.png

marked_bear.png

Answer 1

[Image Analyst]

Like this:

I = imread('bear.png');
J = rgb2gray(imread('marked_bear.png'));
mask = I-J; % Weird, but okay...whatever.
I(mask~=0)=255;

[Ba Ba Black Sheep!]

1. function needs to be there. Coz, it's not that simple always.

2. Why do you think it's weird?

[Image Analyst]

To make a funciton

function I = MaskI(I, J)
mask = I - J;
I(mask ~= 0) = 255;

Usually people use logical operations to create masks that are binary images, not arithmetic ones that produce integer images, that may depend on clipping, and then which still have to use logical operations anyway, such as mask ~=0. Most people, seeing what you intend, would have done:

mask = I ~= J;

and most people who write code for industry that needs to be maintained would use more descriptive variable names than I and J. I know Mathworks doesn't because they want their demo code to look as simple as possible, but I wouldn't let them write code like that in my company.

[Ba Ba Black Sheep!]

I was talking about retaining some_function().

[Image Analyst]

OK, then...

function I = some_function(I, J)
mask = I - J;
I(mask ~= 0) = 255;

There. It's vectorized, still has a function, and that function is called "some_function" (instead of MaskI as I had called it), just like you asked for.