Book Image

Hands-On Image Generation with TensorFlow

By : Soon Yau Cheong
Book Image

Hands-On Image Generation with TensorFlow

By: Soon Yau Cheong

Overview of this book

The emerging field of Generative Adversarial Networks (GANs) has made it possible to generate indistinguishable images from existing datasets. With this hands-on book, you’ll not only develop image generation skills but also gain a solid understanding of the underlying principles. Starting with an introduction to the fundamentals of image generation using TensorFlow, this book covers Variational Autoencoders (VAEs) and GANs. You’ll discover how to build models for different applications as you get to grips with performing face swaps using deepfakes, neural style transfer, image-to-image translation, turning simple images into photorealistic images, and much more. You’ll also understand how and why to construct state-of-the-art deep neural networks using advanced techniques such as spectral normalization and self-attention layer before working with advanced models for face generation and editing. You'll also be introduced to photo restoration, text-to-image synthesis, video retargeting, and neural rendering. Throughout the book, you’ll learn to implement models from scratch in TensorFlow 2.x, including PixelCNN, VAE, DCGAN, WGAN, pix2pix, CycleGAN, StyleGAN, GauGAN, and BigGAN. By the end of this book, you'll be well versed in TensorFlow and be able to implement image generative technologies confidently.
Table of Contents (15 chapters)
Section 1: Fundamentals of Image Generation with TensorFlow
Section 2: Applications of Deep Generative Models
Section 3: Advanced Deep Generative Techniques

Chapter 7: High Fidelity Face Generation

As GANs began to become more stable to train, thanks to improvements to loss functions and normalization techniques, people started to shift their focus to trying to generate higher-resolution images. Previously, most GANs were only capable of generating images up to a resolution of 256x256, and simply adding more upscaling layers to the generator did not help.

In this chapter, we will look at techniques that are capable of generating images of high resolutions of 1024x1024 and beyond. We will start by implementing a seminal GAN known as Progressive GAN, sometimes abbreviated to ProGAN. This was the first GAN that was successful at generating 1024x1024 high-fidelity face portraits. High-fidelity doesn't just mean high-resolution but also a high resemblance to a real face. We can have a high-resolution generated face image, but if it has four eyes, then it isn't high fidelity.

After ProGAN, we will implement StyleGAN, which...