There are two key components to building this network: the discriminator and generator methods. The following graphic shows an illustration of these two networks in action:
The general network architecture for Image-to-Image Translation with Conditional Adversarial Networks
The next two sections will cover the basic pseudocode for each component.