In the previous chapter, we got the first taste of the potential of GANs to learn the connections between latent vectors and generated images and made a vague observation that latent vectors somehow manipulate the attributes of images. In this chapter, we will officially make use of the label and attribute information commonly seen in open datasets to properly establish the bridge between latent vectors and image attributes.
In this chapter, you will learn how to use conditional GANs (CGANs) to generate images based on a given label and how to implement adversarial learning with autoencoders and age human faces from young to old. Following this, you will be shown how to efficiently organize your source code for easy adjustments and extensions.
After reading this chapter, you will have learned both supervised and unsupervised approaches...
