A GAN uses a stack of neural networks to come up with a new image that looks very similar to the original set of images. It has a variety of applications in image generation, and the field of GAN research is progressing very quickly to come up with images that are very hard to distinguish from real ones. In this section, we will understand the basics of a GAN – how it works and the difference in the variations of GANs.
A GAN comprises two networks: a generator and a discriminator. The generator tries to generate an image and the discriminator tries to determine whether the image it is given as an input is a real image or a generated (fake) image.
To gain further intuition, let's assume that a discriminator model tries to classify a picture into a human face image, or not a human face from a dataset that...
