How to find the difference of adjacent pixels in an image? Thanks

How to find the difference of adjacent pixels in an image? Thanks

Answer 1

[Image Analyst]

In which direction? Assuming left and right, you can use conv2() or imfilter():

img = imread('moon.tif');
diffImage = imfilter(img, [1, -1]);
imshow(diffImage, [])

[Jab]

could you please tell me what does the [1,-1] represent?

[Image Analyst]

It says take every pixel and subtract the one to the right of it.

[Jab]

Thank you very much. I got it!

[Sidra Aleem]

I have a 3 X 3 Matrix. I want to subtract adjacent pixels on boundaries to subtract from each other in clockwise manner skipping the central pixel. Example:

A= [ 0,0,1; 1,1,0; 0,0,1]; Subtraction of adjacent pixel at boundaries. (0-0)=0 , (0-1)= -1; (1-0)=1; (0-1)= -1, (1-0)=1, (0,0)=0; (0-1)=-1, (1-0)=1. I tried

    X = [0 0 1;
         1, 1,0;
         0,0,1]; 
  Y = diff(X,1,2);
 But its the difference between columns.
I also tried index wise subtraction. But its not giving correct result.
 Thanks in advance

[Image Analyst]

I'm not sure what you mean by boundaries, like if you want the perimeter of some large, irregularly-shaped blob, or if you just want the 8 closest neighbor pixels. If you want the 8 closest, you can use conv2:

kernel = [-1, -1, -1; -1, 8, -1; -1, -1, -1];
output = conv2(double(X), kernel, 'same');
imshow(output, []);

[Sidra Aleem]

The pixel in the center is my pixel (P) under consideration. The pixels that are present in its 8 neighborhood are considered as the pixels at boundary of P. Now i want to subtract the adjacent pixels at boundaries in clockwise manner. Depending on the resulting value I have to update value of P.

[Sidra Aleem]

Below figure illustrates what I want to do

[Image Analyst]

That would have been helpful to know in advance. I have no time to program this up for you but it's very similar to my local binary pattern demo where I get the values in a circle around the pixel like that. You just have to add or subtract them rather than take them as a binary number. See attached demo.