Deepfake Video Detection using Deep Learning Approach
Aishwarya Jakka, J. Vakula Rani, Manoj Challa, M Vinay Kumar, Gopikrishnan Kookkal · 2024
The rapid advancement of deepfake production technologies is threatening the reliability of media content. Deepfake videos have the potential to increasingly deceive and manipulate viewers by propagating false information, influential public opinion, and damaging reputations. The pseudo-realistic content produced by these methods is exceedingly difficult for conventional detection techniques to identify. Therefore, there is a need for effective deepfake detection and identification of such videos and images. This research employs neural networks, specifically, ResNet, EfficientNet, and InceptionNet, to distinguish between fraudulent and authentic videos and images. FF++ 2020, Deeper-Forensics, and DFDC datasets are used, and evaluated these models Our results demonstrate the potential of these neural networks in countering the growing threat posed by deepfake technology.