IsSwap: Deep Fake Detection

In this era of technology the intake of information through digital media has grown exponentially and it has provided people with personal motives to spread falsified information among the masses to create biased opinions and a sense of unrest. The falsified information is provided to the people especially during elections to create political unrest among the masses or simply to spread a rumour. Since most of the information consumed by people is in the form of videos, it has become a great target spot for people with malicious intent. Deep-Fake is an emerging threat to celebrities, political faces and JSON common people. The paper is aimed at overcoming the given challenge by providing a fast and reliable method to determine the authenticity of a given video. We have proposed a model for detecting deep fake videos via XceptionNet and ResNet50 together with multiple hidden layers of neural networks. The results shows that the proposed method outperformed other state-of-the-art methods in terms of precision and recall rate and performed well in predicting whether the video is fake or real.