Comparative Analysis of Deep Fake Detection Techniques

Abstract

Deep learning and artificial intelligence are important knowledge areas that have provided solutions allowing the successful resolution of complex problems. Some of these problems include, but are not limited to, human-level control, data analytics and other digitisation challenges. One of the offshoots of deep learning is a concept termed ‘deepfake’, which can be described as the imposition of video of a face image from a source to video of the face image of a target individual in order to make the targeted person appear to express the content of the source video [2]. It is important to establish the fact that deepfakes have been used for malicious purposes, becoming a threat to national security, privacy, democracy, and society at large. It is, therefore, fundamental to review the science behind the method, and the available detection techniques to curtail this digital innovation, so as to reduce its level of threat; that is the focus of this paper.