Federal Service for Intellectual Property issued two patents to the bank — each of them represents two different technologies. Sber plans to use them in a joint system, due to which the detection accuracy will greatly increase.
In particular, it will be possible to detect deepfakes in frames with the presence of several persons. In total, based on the test results, the system was able to detect about 98% deepfakes, which is significantly higher than that of existing analogues.
Detection of deepfakes is a serious breakthrough in the field of security, which clearly demonstrates Sber’s IT focus. The system will allow:
1. Protect yourself from cyber attacks by bypassing Face Recognition and Liveness Detection
2. Ensure security when calling via videoconferencing
3. Fight fakes more effectively by identifying information attacks at an early stage
Here is how Stanislav Kuznetsov, deputy chairman of the board of Sber, commented on the patents: “Sber is constantly improving its cybersecurity systems and conducting scientific research in this area. Our patented inventions will allow us to effectively solve problems of identifying information attacks and cyber attacks using deepfake technologies. Synthetic videos used to attack Sberbank and our clients will be identified at the earliest stages.”
Sber has patented technologies that can recognize deepfakes
Rospatent issued two patents to Sber Bank for technologies that help identify deepfakes. According to company representatives, the system’s efficiency is 98%. This is higher than existing analogues.
“ВЧК ОГПУ”
Specialization: International Criminal Investigations Description: Anna is a journalist with a global outlook, specializing in uncovering international criminal networks, including human and arms trafficking. Her work involves in-depth analysis of transnational crime and writing comprehensive investigative articles.
Soledad is the Best Newspaper and Magazine WordPress Theme with tons of options and demos ready to import. This theme is perfect for blogs and excellent for online stores, news, magazine or review sites.