Vulnerability assessment and detection of Deepfake videos

Recent advances in automated video and audio editing tools, Artificial Intelligence (Generative Adversarial Networks), and social media allow creation and fast dissemination of high quality tampered video content. Such content already led to appearance of deliberate misinformation, coined `fake news', which is impacting political landscapes of several countries. A recent surge of videos, often obscene, in which a face can be swapped with someone else's using GANs, so called Deepfakes, are of a great public concern. As the technology for generating such synthetic video content, including swapped faces, is constantly improving, the development of efficient tools that can automatically detect these videos is of a paramount importance. The objective of this project is (1) to assess how realistic the generated Deepfake videos to fool both human observers and automatic recognition systems, and (2) to develop effective mechanisms to detect Deepfakes.
Hasler Stiftung (Hasler Foundation)
Aug 01, 2019
Jul 31, 2020