Introducing Pix2Pix with pseudocode

When image-to-image translation with conditional adversarial networks, released as Pix2Pix, came out in 2016, it was widely praised as a simple style-transfer network that works out of the box. The network requires less parameter tuning than other techniques in the field, and you'll see the power of this network at the end of this chapter. In this recipe, we're going to cover the basics of the algorithm and present pseudocode prior to implementing the actual code.

Getting ready

Grab the Image-to-Image Translation with Conditional Adversarial Networkspaper: https://arxiv.org/pdf/1611.07004.pdf.

Then, read it and move on—we'll cover the basics of implementing the paper in this recipe.

How to do it...

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...