In this chapter, we will be exploring the convolution neural network internals by visualizing what these layers are trying to learn. Then, we will look at the neural style transfer problem, and how to solve that problem by using the knowledge about the layers learning process. We will then continue with a content cost function intuition, together with a bit more formal mathematical definition of the cost and derivation. We will go through the details of building the style cost function, which are slightly more complex, and how to efficiently capture the style of an image in terms of convolution layers. Finally, we will present the optimized architecture, together with the core details for the Java implementation for neural style transfer. We will also show an image sample for a few iterations.

We will be covering the following topics:

• What are convolution network layers learning
• Neural style transfer
• Applying content cost function
• Applying style cost function
• Building a neural network that produces art