3D Generative Adversarial Networks

3D Generative Adversarial Networks

Generative Adversarial Networks (GANs) are a leading deep generative model that use deep neural networks (DNNs) to learn from a set of training examples in order to create new design instances with a degree of flexibility and fidelity that outperform competing generative approaches. GANs have been used in the creation of 2D designs but their application to 3D synthesis problems has been less explored. This research attempts to fill this gap and explores their use in the synthesis of building forms. Specifically, a set of 700 high rise buildings from downtown NYC is used to train a 3D GAN. The GAN is then used to produce new designs. Some examples from this synthesis task can be seen in Figure 2.