Applied Computer Vision through Artificial Intelligence

2Generative Adversarial Networks: Theory and Application in Synthesis

Manoj Kumar Pandey¹, Priyanka Gupta², Triveni Lal Pal³* and Ayush Kumar Agrawal⁴

¹Dept. of CSE, Chandigarh University, Mohali, India

²Department of CSIT, Guru Ghasidas Central University, India

³Dept. of CSE, Galgotias University, Greater Noida, India

⁴Dept. of IT & CS, Dr. C. V. Raman University, Bilaspur, India

Abstract

Generative adversarial networks (GANs) have gained prominence as a ground-breaking technology with a significant role in both theory and practical applications, particularly in the domain of synthesis. GAN has emerged as a powerful tool for synthesis generation for images and found its extensive application for various domains. Two models are associated with the GAN for unseen data generation i.e. generator and discriminator. In general, GANs have transformed the domain of generative modeling. GANs have transformed the domain of generative modeling in general. By conceptualizing generative model training as a game between two networks, they present a novel method for doing so but the discriminator network calculates the differences in genuine and generated data, and the generator network generate the some of the data looked as real data. Realistic data is produced as a result of this adversarial process, which also helps both networks get better. In practical applications, GANs have found extensive use in various areas like NLP, computer vision and healthcare. The chapter systematically ...