A Survey on Deepfake Video Detection Techniques Using Deep Learning

Abstract

Deep learning is an effective technology that has been widely used in different ways. ‘DeepFake’ videos are generated using deep learning technology called generative adversarial network where the videos are created with swaped faces in a video, altered facial expressions, change of gender, creation of fake video content and altered facial features. Fake videos are used in the situations like extortion of money, terrorism events or create political agitation. The deepfake technologies are used for positive purposes, such as film-making and virtual reality. The good quality results from deepfakes are very hard to recognised with people eyes. Many deep learning techniques are built to detect deepfake content in images and videos. A comparative study of the performance of various deepfake videos detection models in the deep learning is preseted in this paper. Convolutional Neural Network (CNN) models include ResNet, VGG16, Efficientnet etc.. Recurrent Neural Network (RNN) model includes Long Short-Term Memory LSTM.