Deepfake Detection System Using Deep Neural Networks

Abstract

The rapid progress in technology and automation has enabled sophisticated manipulation of multimedia content, blurring the line between real and fabricated media. Deepfake technology, driven by deep learning and Generative Adversarial Networks (GANs), creates hyper-realistic fake content, with applications spanning video games, films, and advertising. However, this technology also carries substantial societal risks, fostering misinformation and explicit content. To mitigate these concerns, this paper presents a Deepfake detection system that utilizes deep neural networks to discern genuine from forged images. Frames are extracted from videos and face detection and face cropped are performed. LSTM and ResNext CNN are utilized to generate a feature vector. The proposed system uses the Anvil platform to design the front end and Visual Studio and Jupyter Notebook for the back end. A publicly available dataset was used to train and test the model. The proposed model achieved an impressive 86% accuracy on video dataset.