It really works by initial capturing cross-element map correlations after which you can spatial correlations. This enables more efficient use of model parameters.top) from the input volume for another convolutional layer. The pooling layer would not impact the depth dimension of the volume. The operation carried out by this layer is also referred t… Read More