doubt on how to divide by using the intensity of an image

hai

i am studying about optic disc localization my work is to convert a two dimensional retinal image into two one dimensional signals the input i have used is a retinal image from drive database for that i am using an algorithm

step 1 calculate the vertical and horizontal edge maps (ev,eh)using the mask[1 0 -1] and its transpose on the green component of the image step2 calculate the edgediff and edgesum which is ev-eh and ev+eh step3 next divide the edgediff by the image intensity fmap1=edgediff(x,y)/I(x,y) my doubt is how to obtain the intensity value

please help me my code is this

g=maskedRgbImage(:,:,2);
imv=conv2(double(g),[1 0 -1],'same');
figure,image((imv));
title ('Vertical Image');
imh=conv2(double(g),[1;0;-1],'same');
figure,image((imh));
title ('Horizontal Image');
ev=abs(imv);
eh=abs(imh);
edgediff=imsubtract(ev,eh);

next step is calculating the image intensity and

fmap1=imdivide(edgediff,original image)

i have divided by using the original image is it correct how to calculate the image intensity and divide thanks in advance please help me

Answer 1

[David Young]

You could try using rgb2gray on the initial image. Looking at your code, this would be:

I = rgb2gray(maskedRgbImage);

then

fmap1 = imdivide(edgediff, I);

I am not certain that this is what is intended, however. The instructions do not seem to correspond to a standard algorithm: it seems a little strange to compute the differences based on the green channel alone, but then to divide by the total intensity. Another possibility is that you are simply meant to divide by g. (The purpose of subtracting the horizontal differences from the vertical differences is also hard to understand.)

If at all possible, go back to your instructor (or whoever gave you the instructions) and request clarification.

[ramya raj]

thanks for your immediate response

actually i am working with the journal paper to locate the optic disc and the algorithm is as follows

1. Define a sliding window with dimensions (image

height, 2 main vessel width)

2. Get an image of horizontal edges (EH) and an image

of vertical edges (EV)

3. Calculate EdgeDiff = |EH| - |EV|; where |.| is the

absolute operator

4. Slide the window over the image from right to left

and at each location,

i) Calculate F = sum of EdgeDiff

ii) Calculate G = sum of pixels’ intensity values

iii) Calculate the ratio Hposition = F / G

5. The horizontal location of the optic disc (OD_H) is

the location of the maximum value in Hposition

this is what i am doing

i have calculate the edge difference and i don't know how to divide by using the intensity.

[David Young]

Without implementing the system myself from the journal paper, I don't think I can offer more advice. My suggestion above about how to do the division still stands.