International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 03 | May 2023
p-ISSN: 2395-0072
www.irjet.net
Exploring The Potential of Generative Adversarial Network: A Comparative Study Of GAN Harsh Shah 1, Kaustubh Kabra2, Onasvee Banarse3, Akash Mete4 1,2,3,4 Student, Department of Computer Engineering,
All India Shri Shivaji Memorial Society’s Institute of Information Technology, Pune. India. ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Generative Adversarial Networks (GANs), a class
improving its ability to distinguish between fake and real data. The discriminator network learns to discriminate between the fake data and the actual data, while the generator network learns to produce artificial data that is comparable to the training data. Both networks develop over time because of this competitive dynamic, producing highquality data that is nearly indistinguishable from real data.
of deep learning models that creates new data samples that resemble the original data, are in-depth examined in this research study. The article covers many GAN subtypes, including vanilla GANs, MedGANs, StyleGANs, and CycleGANs, and analyses their designs and training approaches. The study examines the many GAN applications, including text-to-image synthesis, data augmentation, and picture and video creation. There is also discussion of the difficulties each type of GAN method faces, including mode collapse, instability, and vanishing gradients.
In this review paper, we present a comprehensive analysis of GANs and their various types, including VanillaGAN, StyleGAN, CycleGAN, and MedGAN. We begin by providing an overview of the GAN architecture and working principle, followed by a literature survey of various GAN models used in different fields. For each GAN type, we discuss their architecture, working principle, applications, related work, challenges, and future directions. Additionally, we provide a comparative analysis of these GAN models based on their performance, advantages, and limitations.
In-depth analysis is also given to the technical features of GANs, including the generator and discriminator networks, training loss functions, and regularization techniques. The research study examines current advancements in GANs, including self-attention, adversarial autoencoders, and attention mechanisms. Additionally, the paper addresses the ethical issues related to GANs, such as the possible exploitation of data created by GANs and bias in training data.
Furthermore, we discuss the current state of research in the field of GANs and highlight some of the recent developments in GANs, including attention mechanisms, progressive growing, and disentangled representations. Finally, we conclude the review paper with future research directions in GANs, highlighting potential areas for improvement and the challenges that need to be addressed for the effective implementation of GANs in real-world applications.
The future potential and developments of GANs are discussed in the study, including its use to unsupervised representation learning and the creation of novel GAN architectures. The study emphasizes the need for more study to overcome GANs' problems and broaden their application to other fields. GANs are a fast-developing subject of study with enormous potential in many areas.
GANs work on the principle of adversarial training, in which two neural networks, the generator and the discriminator, compete in a two-player minimax game. The generator network takes a random input and generates synthetic data, while the discriminator network tries to distinguish between the synthetic data and real data. The two networks are trained simultaneously, with the generator attempting to generate synthetic data that can fool the discriminator into believing that it is real, while the discriminator attempts to accurately classify the real and synthetic data. Through this adversarial training process, the generator network learns to generate synthetic data that closely resembles the real data. This approach has shown remarkable success in generating high-quality synthetic images and has since been extended to other types of data, such as audio, video, and text.
Key Words: Generative Adversarial Networks, GAN architectures, GAN applications, Computer Vision, Anomaly Detection.
1. INTRODUCTION Generative Adversarial Networks (GANs) have gained significant attention in recent years for their ability to generate realistic data in various fields such as computer vision, natural language processing, and healthcare. GANs are composed of two neural networks: a generator network that creates fake data and a discriminator network that distinguishes between the generated fake data and real data. The two networks compete against each other, with the generator network attempting to produce data that can fool the discriminator network, and the discriminator network
© 2023, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 1331