So far, we've learned how to synthesize images, text, and audio with GANs. Now, it's time to explore the 3D world and learn how to use GANs to create convincing 3D models.
In this chapter, you will learn how 3D objects are represented in computer graphics (CG). We will also look into the fundamental concepts of CG, including camera and projection matrices. By the end of this chapter, you will have learned how to create and train 3D_GAN to generate a point cloud of 3D objects, such as chairs.
You will know the fundamental knowledge of the representation of 3D objects and the basic concept of 3D convolution. Then, you will learn to construct a 3D-GAN model by 3D convolutions and train it to generate 3D objects. You will also get familiar with PrGAN, a model that generates 3D objects based on their black-and-white 2D views.
The following...
