Deep Fake Face Detection Using Machine Learning and Various GAN Models

Abstract

Remarkable advancements in Deep learning, incredibly genuine AI generated pretended videos have been created. To appropriately construct the Deep fake phenomenon, new Deep fake detection algorithms must be designed for the exploitation of this powerful A.I. technology has a major effect on person lives. However, traces left by Generative Adversarial Network (GAN) engines during the construction of Deep fakes can be discovered by studying ad-hoc frequencies. We'll utilize the Deep Convolutional Generative Adversarial Network (DCGAN) and Style GAN, which have both shown to be quite good at creating images. We covered the theoretical aspects of GAN as well as our methods for developing a DCGAN Model using MNIST and CelebA data sets.