This paper describes how watermarking technology can be used to prevent the proliferation of Deepfake news. In the proposed system, digital watermarks are embedded in the audio and video tracks of video clips of trusted news agencies at the time the videos are captured or before they are distributed. The watermarks are detected at the social media network’s portals, nodes, and back ends. The embedded watermark imparts a unique identifier to the video, that links it to a blockchain. The watermarks also allow video source tracking, integrity verification, and alteration localization. The watermark detectors can be standalone software applications, or they can be integrated with other applications. They are used to perform three main tasks: (1) they alert the internet user when he watches an inauthentic news video, so that he may discard it, (2) they prevent a Deepfake news video from propagating through the network (3) they perform forensic analysis to help track and remove Deepfake news video postings. The paper includes Proof-of- Concept simulation results.
The advent of Generative Adversarial Networks (GANs) has brought about completely novel ways of transforming and manipulating pixels in digital images. GAN based techniques such as Image-to-Image translations, DeepFakes, and other automated methods have become increasingly popular in creating fake images. In this paper, we propose a novel approach to detect GAN generated fake images using a combination of co-occurrence matrices and deep learning. We extract co-occurrence matrices on three color channels in the pixel domain and train a model using a deep convolutional neural network (CNN) framework. Experimental results on two diverse and challenging GAN datasets comprising more than 56,000 images based on unpaired image-to-image translations (cycleGAN [1]) and facial attributes/expressions (StarGAN [2]) show that our approach is promising and achieves more than 99% classification accuracy in both datasets. Further, our approach also generalizes well and achieves good results when trained on one dataset and tested on the other.