April 2018
Beginner to intermediate
282 pages
6h 52m
English
The pooling layer further reduces the size of the feature representation by applying a pooling function. There are different kinds of pooling functions, such as average, min, and max. Max pooling is widely used as it tends to keep the max values of a feature map for each stride. This is similar to the convolution layer where we have a sliding window and the window slides over the feature map to find the max value within each stride. The window size in a pooling layer is typically less than that used in the convolution layer.
The pooled feature map is then flattened to a 1D representation to be used in a fully connected layer.