EP-4738847-A1 - VISUALLY EMBEDDING AUTHENTICITY PROOF DATA IN VIDEO CONTENT

EP4738847A1EP 4738847 A1EP4738847 A1EP 4738847A1EP-4738847-A1

Abstract

Authenticity proof data can be visually embedded in a video file or stream. The visual representation of the authenticity proof data is perceivable to users and scannable by user devices for validating the authenticity proof data. This allows for verification of content integrity through direct observation or scanning, enhancing trust and security in digital media distribution.

Inventors

TAYLOR, SAMUEL
MARTIN, Anja

Assignees

PlayAI GmbH

Dates

Publication Date: 20260506
Application Date: 20241101

Claims (15)

A video file or video stream comprising: video content; and a visual representation of authenticity proof data visually embedded in the video content; wherein the visual representation of the authenticity proof data is visually perceivable to a user; and wherein the visual representation of the authenticity proof data is scannable by a user device to validate the authenticity proof data.
The video file or the video stream of claim 1, wherein the visual representation of the authenticity proof data comprises a user-recognizable graphical element.
The video file or the video stream of claim 2, wherein the graphical element comprises a plurality of segments, wherein each segment of the plurality of segments is displayed in a color of a color palette.
The video file or the video stream of claim 3, wherein the color palette is associated with a base color, and wherein each color of the color palette is a shade of the base color.
The video file or the video stream of claim 3 or 4, wherein each color of the color palette represents a character in an encoding alphabet.
The video file or the video stream of claim 5, wherein the colors of the color palette are selected as contrasting colors for common combinations of characters.
The video file or the video stream of any one of claims 2 to 6, wherein the graphical element is an animated graphical element, wherein the animation comprises changing the colors of the segments.
The video file or the video stream of any one of claims 1 to 7, wherein the authenticity proof data comprises a message; wherein the message comprises one or more of: a video summary; a product name; a version; a date; an instance identifier; a link reference; padding information.
The video file or the video stream of any one of claims 1 to 8, wherein the authenticity proof data comprises a digital signature.
The video file or the video stream of any one of claims 5 to 9, wherein: the video content comprises AI-generated video content, in particular an avatar of a real-world athlete; the plurality of segments consists of eight segments; the encoding alphabet is the hexadecimal alphabet.
The video file or the video stream of claim 10, wherein each segment pair encodes a Base64 character using two hexadecimal values.
A method of encoding authenticity proof data in a video file or video stream according to any one of claims 1 to 11, the method comprising: obtaining a visual representation of the authenticity proof data; and visually embedding the visual representation of the authenticity proof data into video content of the video file or the video stream.
A method of decoding authenticity proof data encoded in a video file or video stream according to any one of claims 1 to 11, the method comprising: observing a visual representation of the authenticity proof data visually embedded in video content of the video file or the video stream; and validating the authenticity proof data.
A data processing apparatus comprising means for carrying out the method of claim 12 and/or claim 13.
A computer program, or a computer-readable medium storing a computer program, the computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of claim 12 and/or claim 13.

Description

TECHNICAL FIELD This disclosure relates to the field of digital content protection, specifically to methods and systems for authenticating content, in particular video content. This disclosure also relates to the field of artificial intelligence (Al) and machine learning (ML), specifically to applications where Al-generated content needs to be authenticated. BACKGROUND The proliferation of Artificial Intelligence (Al) generated "deepfake" videos has led to concerns about their authenticity and potential misuse. Al-generated videos are becoming increasingly difficult to distinguish from real-life recordings, making it challenging to distinguish between authentic and manipulated recordings. This compromise can be exploited by malicious actors seeking to disseminate false or misleading information. When the subject matter involves well-known or recognizable individuals ("very important persons"; VIPs), such as celebrities, athletes, politicians, or business leaders, the lack of robust authentication mechanisms can lead to confusion and misinformation. The absence of authorized or licensed products featuring these VIPs means that consumers have no way of determining if a particular video is an official representation of the person. Furthermore, with the rise of social media platforms and online sharing, videos are often captured, recorded, and shared without any guarantee of being generated in a controlled environment. This raises concerns about the integrity and authenticity of these videos, as well as their potential impact on public perception and reputation. This undermines trust in video content and creates opportunities for malicious actors to spread disinformation, propaganda, or other forms of malicious activity. In light of these challenges, there is a need to develop effective techniques that ensure the authenticity and integrity of video content, in particular Al-generated videos, thereby mitigating the security risks associated with their dissemination. It is therefore an objective of the present disclosure to provide techniques that ensure the authenticity of video content and provide end-consumers with an easy way to verify its legitimacy, thereby overcoming the drawbacks of the prior art at least in part. SUMMARY OF THE DISCLOSURE The above and other objectives may be achieved by the subject-matter defined by the independent claims. Advantageous modifications of embodiments of the present disclosure are defined in the dependent claims as well as in the description and the drawings. One aspect of the present disclosure provides a video, such as a video file or video stream, and/or a data structure encoding such a video, video file or video stream. The video file or the video stream may comprise video content. Although the aspects and embodiments of the present disclosure will be explained in connection with a video file or a video stream, they can be applied in connection with any data structure configured to store video content. The video file or the video stream may comprise a visual representation of authenticity proof data. It may be provided that the visual representation of the authenticity proof data is embedded, in particular visually embedded, in the video content. It may be provided that the visual representation of the authenticity proof data is visually perceivable to a user. Embedding authenticity proof data directly into the video content offers several significant advantages, making it an effective solution for ensuring the authenticity, integrity and legitimacy of digital videos. Firstly, this approach provides a way to convey the authenticity proof data that is agnostic of where the video is hosted. This means that regardless of whether the video is stored on a social media platform, a website, or a physical device, the embedded authenticity proof remains intact and accessible. The viewer can verify the authenticity of the video without relying on any specific hosting environment or software. Secondly, this method is not reliant on built-in software. Unlike traditional approaches that require specialized software to detect and authenticate videos, embedding authenticity proof data in the video content itself eliminates the need for additional tools or plugins. This makes it a more accessible and user-friendly solution for verifying video authenticity. Thirdly, the embedded authenticity proof data is tolerant of being copied and/or transcribed. When a video is shared, recorded, or transcribed, the authenticity proof remains intact, ensuring that the verification process can be applied to any subsequent copies or versions of the original content. Fourthly, by embedding the authenticity proof data directly into the video content, it becomes an integral part of the video itself. This means that whenever the video is copied, recorded, shared, etc., the digital signature goes with it. The authenticity proof remains linked to the video, ensuring that any attempts to manipulate or alter t