Consider the network in Figure $4,$ which is constructed by two $55$ convolution kernels. What is the receptive field of one pixel in layer $2?$ Please draw a $2$D graph $($excluding the channel dimension$)$ to illustrate your calculation. Note: The cases where the convolution kernels are on the edge of the feature map can be ignored in this question. $2.$ Receptive Fields

$($a$)$ Consider the network in Figure $4,$ which is constructed by two $\backslash (5\backslash$times $5\backslash )$ convolution kernels. What is the receptive field of one pixel in layer $2?$ Please draw a $2$D graph $($excluding the channel dimension$)$ to illustrate your calculation. Note: The cases where the convolution kernels are on the edge of the feature map can be ignored in this question.

Figure $4$: Convolutional Architecture

$($b$)$ Now consider adding a $\backslash ($ n $\backslash )-$by$-\backslash ($ n $\backslash )$ max$-$pooling layer following layer $2.$ What will be the receptive field after the max$-$pooling layer? Note: The cases where the pooling kernel is on the edge of the feature map can be ignored in this question.