Researchers use machine learning to quickly detect video face swaps

We all know that AI can be used to swap faces in photos and videos. People have, of course, taken advantage of this tool for some disturbing uses, including face-swapping people into pornographic videos — the ultimate revenge porn. But if AI can be used to face swap, can’t it also be used to detect when such a practice occurs? According to a new paper on arXiv.org, a new algorithm promises to do just that, identifying forged videos as soon as they are posted online.

The team, led by Andreas Rossler at the Technical University of Munich, developed machine learning that is able to automatically detect when videos are face swapped. They trained the algorithm using a large set of face swaps that they made themselves, creating the largest database of these kind of images available. They then trained the algorithm, called XceptionNet, to detect the face swaps.

XceptionNet clearly outperforms its rival techniques in detecting this kind of fake video, but it also actually improves the quality of the forgeries. Rossler’s team can use the biggest hallmarks of a face swap to make the manipulation more seamless. It doesn’t fool XceptionNet, but in the long run, it could make it harder for other methods to detect faked videos.

Via: MIT News

Source: arXiv.org

Post Author: martin

Avatar
Martin is an enthusiastic programmer, a webdeveloper and a young entrepreneur. He is intereted into computers for a long time. In the age of 10 he has programmed his first website and since then he has been working on web technologies until now. He is the Founder and Editor-in-Chief of BriefNews.eu and PCHealthBoost.info Online Magazines. His colleagues appreciate him as a passionate workhorse, a fan of new technologies, an eternal optimist and a dreamer, but especially the soul of the team for whom he can do anything in the world.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.