City-GAN: Learning architectural styles using a custom Conditional GAN architecture
This work addresses a domain-specific problem in architectural style generation, offering incremental improvements for applications in urban planning or design.
The paper tackles the problem of generating realistic building images for major cities using GANs, proposing a custom Conditional GAN architecture that shows superior performance compared to existing methods, as verified through extensive experiments.
Generative Adversarial Networks (GANs) are a well-known technique that is trained on samples (e.g. pictures of fruits) and which after training is able to generate realistic new samples. Conditional GANs (CGANs) additionally provide label information for subclasses (e.g. apple, orange, pear) which enables the GAN to learn more easily and increase the quality of its output samples. We use GANs to learn architectural features of major cities and to generate images of buildings which do not exist. We show that currently available GAN and CGAN architectures are unsuited for this task and propose a custom architecture and demonstrate that our architecture has superior performance for this task and verify its capabilities with extensive experiments.