Generating Realistic Marble Textures using Generative Adversarial Networks

This bachelor thesis explores the application of Generative Adversarial Networks (GANs) in generating realistic marble images, with a focus on the manufacturing industry. The study is conducted at Breton S.p.A, a leading company in the production of machines and technologies for processing natural stone, glass, and metal. The research involves implementing various GAN models, including DCGAN, WGAN, and StyleGAN, and evaluating their performance in terms of image quality and similarity to real marble. The results demonstrate the potential of GANs in generating high-quality marble images that can be used for product design and prototyping.