Exploring Generative Adversarial Networks (GANs) for Deepfake Detection: A Systematic Literature Review

Abstract

This paper explores the application of Generative Adversarial Networks, also known as GANs for deepfake detection. It investigates the use of various GAN models in deepfake detection, subjecting them through training against other models, and in many different datasets. Through that process, this paper aims to improve detection performance and adaptability to many evolving deepfake techniques. Evaluations are conducted on benchmark deepfake datasets, evaluation metrics, and comparation between the proposed GAN-based detection methods with the most prominent approaches. The results provide insights to the strengths and weaknesses of GANs for deepfake detection and help to develop more robust systems in the same field.