How to compute the horizontal and vertical gradients of a gray-scale image without the use of imgradientxy()

Hi guys, I'm completely new to matlab and am having some troubles with getting a seemingly trivial part of an assignment. I need to be able to find the horizontal and vertical gradients of a gray-scale image without the use of matlab functions such as imgradientxy(). The kernal I am using to find gradient is (-1,0,1). My attempt was to use loops to loop through the location of each pixel and change them one by one like so:

original = imread('grayscale.jpg')
[x,y] = size(original)

hor_grad = zeros(x,y)
ver_grad = zeros(x,y)

for i=1:x
   for j=2:y-1
  hor_grad(i,j) = -1*original(i,j-1)+original(i,j+1)
   end
end

for i=2:x-1
   for j=1:y
  ver_grad(i,j) = -1*original(i-1,j)+original(i+1,j)
   end
end

imshow(hor_grad)
imshow(ver_grad)

However I quickly learned that this approach will not work. I am at a loss here. What is a concise way to do this?

Answer 1

[Arun Mathamkode]

Although this is not the best way of doing it, should also work fine. Please convert the original image to double before processing to avoid errors due to integer precision.

    original = imread('grayscale.jpg');
    original=double(original);

    [x,y] = size(original);
    hor_grad = zeros(x,y);
    ver_grad = zeros(x,y);
    for i=1:x
       for j=2:y-1
      hor_grad(i,j) = -1*original(i,j-1)+original(i,j+1);
       end
    end
    for i=2:x-1
       for j=1:y
      ver_grad(i,j) = -1*original(i-1,j)+original(i+1,j);
       end
    end
    figure;imshow(hor_grad,[])
    figure;imshow(ver_grad,[])