Generative Adversarial Networks: A Comprehensive Review

Numerous models in the deep learning field have been created as a result of the rise in processing capacity. A generative model called Generative Adversarial Networks (GAN) first appeared in 2014. Many architectures of GAN have been proposed in the process of research conducted on GAN. Any GAN architecture is the result of the competition between two networks, the Generator and Discriminator, to determine the distribution of the sampled data. This process helps in many applications like text to image conversion, style transfer, generating new images, attribute transfer, photo enhancement. This paper helps in knowing about the main working principle of any GAN architecture, recent advances, and applications of GAN.