In this chapter, you will learn to use convolutional neural networks (CNNs) for image recognition. Convolutional neural networks are a variation of neural networks that are particularly well-suited to image recognition because they take into account the relationship between data points in space.
We will cover how convolutional neural networks differ from the basic feedforward, fully connected neural network that we created in the last chapter. The main difference is that the hidden layers in a CNN are not all fully connected dense layers—CNNs include a number of special layers. One of these is the convolutional layer, which convolves a filter around the image space. The other special layer is a pooling layer, which reduces the size of the input and only persists particular values. We will go into more depth on these layers later in the chapter...
