Applications of GANs in Image Generation and Style Transfer
B. Santhosh, Kapinaiah Viswanath · Advances in computational intelligence and robotics book series · 2025
Generative Adversarial Networks (GANs), introduced by Ian Good fellow in 2014, have revolutionized the field of artificial intelligence, particularly in image generation and style transfer. GANs consist of two neural networks, a generator and a discriminator, which are trained simultaneously through adversarial processes. GANs have found extensive applications in various domains of image generation, such as creating high-resolution images, generating images from textual descriptions, and augmenting datasets for machine learning tasks. In style transfer, GANs have enabled the seamless merging of different artistic styles with content images, producing visually appealing and artistically sophisticated results. Techniques like CycleGAN and StyleGAN have pushed the boundaries, allowing for unpaired image-to-image translation and fine-grained control over generated images' style and content. This abstract explores the underlying principles of GANs, highlights key advancements, and discusses their transformative impact on image generation and style transfer.