Convolution operation is defined as follows: f[x, y] *g[x, y] = finanzl-g|x - my-n2] n=-00 n-00 The image patch is defined as follows f= The convolution kernel is defined as 1 2 3 1 1 1 2 3 2 2 1 2 3 3 3 1 2 3 2 2 2 3 1 1/ 1 g= -1 5 -1 0 -1 0 Please evaluate the convolution results for the 2 following cases: (1) Stride is equal to 1, and zero padding is used. (2) What are the results after ReLU? (3) Stride is equal to 2, and zero padding is used.