Machine Learning with Apache Spark Quick Start Guide

Fully connected layer

After the 3-D input data has been transformed through a series of convolution and pooling layers, a fully connected layer flattens the feature maps output by the last convolution and pooling layer into a long 1-D feature vector, which is then used as the input data for a regular ANN in which all of the neurons in each layer are connected to all of the neurons in the previous layer.