how to calculate intensity

I really appreciate your help and I would be so grateful if you help me with this?,

how to calculate, the difference between the number of pixels in a block and the ratio of the sum of pixel intensities in the block to the maximum in the block

i wanted to substitute the values in the equation in the link below.....

http://img89.imageshack.us/img89/1788/bdipt.png

please do reply.....

Answer 1

[Walter Roberson]

I = rgb2gray( im2double( YourImage) );

T = M(x : x+M-1, y : y+M-1);
sum(T(:))
max(T(:))

and I'll leave you to work out the ratio.

[Elysi Cochin]

thank u sir.... sir what does that x and y value denote.... is it row and coulmn..... please reply sir.....

[Walter Roberson]

It denotes the same thing that is in the formula.

[Image Analyst]

This is a common source of confusion. In the formula they have I(x,y) where x is the horizontal index. But in MATLAB, with arrays, the first index is the vertical index. So it's something you always have to watch out for. Fortunately since you're going over all pixels, it doesn't really matter except when you get to the end of a row or columns. You need to calculate

[rows columns numberOfColorChannels] = size(I);

then the first index (what they call x) must go from 1 to rows, not 1 to columns (which you'd think would be the last x value going by the name "x").

[Elysi Cochin]

sir, when i did like this.... i'm getting error... this is what i did....

    [filename pathname] = uigetfile('*.png','Select An Image');
    yourImage = imread([pathname filename]);

    I = rgb2gray( im2double( yourImage ) );
    [x y] = size(I);
    T = M(x : x+M-1, y : y+M-1);
    sum(T(:))
    max(T(:))

This is the error i'm getting.... how to rectify it sir.....

    ??? Index exceeds matrix dimensions.

    Error in ==> Untitled5 at 9
    T = M(x : x+M-1, y : y+M-1);

please do reply sir.....

[Walter Roberson]

The line

T = M(x : x+M-1, y : y+M-1);

should have been

T = I(x : x+M-1, y : y+M-1);

[Elysi Cochin]

sir i changed the above line..... but its still showing that same error.....

[Walter Roberson]

If your blocks are to slide G pixels at a time, e.g., G = 1

for x = 1 : G : size(I,1) - M + 1
  for y = 1 : G : size(I,2) - M + 1
    B = M(x : x+M-1, y : y+M-1);
    P = sum(B(:));
    Q = max(B(:));
    ratio = .... a function of P and Q
  end
end

[Elysi Cochin]

thank u sir...

[Image Analyst]

Of course you could do this all in one to three lines with a vectorized approach using conv2() and imdilate(), like this (untested):

blurredImage = -conv2(yourImage, ones(M)/M^2, 'same') + M^2;
maxImage = imdilate(yourImage, true(M));
BDIP = blurredImage ./ maxImage;