Deep fake Video Detection Using Unsupervised Learning Models: Review

The recent decade has been the most emerging for technical advancements in computer and communication fields. These advancements have made the task of automation much cheaper and available to every other individual at minimum or free of cost. Deep Fakes are one such advancement in the field of image processing. Deep fakes are images or videos developed using deep learning models to create fake images or videos. Although there are advantages of these deep fakes like visual training, powerful simulation, etc... There are also demerits or disadvantages highly possible with these technological advancements such as creating content that never existed, making a person speak something which one has not spoken at all. These mentioned possibilities lead to person identification theft, spreading of false or misinformation, etc... This paper is an attempt to investigate the proposed detection techniques for deep fakes. It focuses on unsupervised deep fake detection techniques and the role of ensemble models for detection.