2d convolution returns empty result

I have an image matrix of size 512x512x3, for each pixel I would like to compute its weight as the average of the neighbouring pixels like below:

weight(y,x,:) = 0.25 * (I(y+1,x,:) + I(y-1,x,:) + I(y,x+1,:) + I(y,x-1,:))

I used the matlab conv2 function to avoid for loops, i.e.

kernel = [0,1,0;
          1,0,1;
          0,1,0]
weight = conv2(kernel, I)

As matlab uses zero padding for the convolution result, I would get the result of size 514x514. But I only want the parts that don't use zero padding:

weight = conv(kernel, I, 'valid')

Unfortunately, this returns empty result. Does anyone know how to solve this problem?

Thank you very much.

Answer 1

[Ameer Hamza]

For 'valid' option, reverse the order of I and kernel

I2 = conv2(I, kernel, 'valid');