Client-Side Video Watermarking via DRM Compositor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital watermarking methods are ineffective in identifying the source of pirated video content at a granular level, such as an individual user account, due to scalability issues and susceptibility to collusion that results in false positives and negatives.
Innovation Solution
Introducing digital watermarks on the client side using compositors in the DRM trust zones of devices, overlaying identifiers on decoded video frames in a way that is not readily detectable to the human eye, allowing for the identification of the source of pirated content by comparing target content with the original source content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If back-end watermarking is used to watermark each variant of each title with two unique watermarks, then the ability to identify end user accounts is improved, but the content storage requirement doubles and the system becomes susceptible to collusion attacks
Solution Approach 1:
The patent inverts the traditional back-end watermarking approach by implementing client-side watermarking. Instead of watermarking content before distribution, the system watermarks content after decryption on the client device. This inversion allows unique identification of individual user accounts without duplicating storage requirements, as each client independently embeds their own identifier in the decoded video frames.
Solution Approach 2:
The patent introduces a compositor as an intermediary component that operates within the DRM trust zone. This compositor takes decoded video frames and embeds unique client identifiers through compositing operations, serving as a mediator between the decryption process and final video output. This intermediary approach enables granular identification without requiring storage duplication.
2Measurement precision
If back-end watermarking with two unique watermarks per variant is implemented, then individual account identification is improved, but the system becomes susceptible to collusion attacks resulting in false positives and negatives
Solution Approach 1:
By inverting the watermarking location from back-end to client-side, the system ensures that each client embeds their own unique identifier independently after decryption. This makes collusion attacks ineffective because even if multiple clients coordinate, each one independently embeds their own identifier, preventing the mixing and matching that enables false positives and negatives in back-end approaches.
Solution Approach 2:
Each client device independently performs the watermarking operation on its own decoded video frames, embedding its own unique identifier. This self-service approach ensures that watermarking is decentralized and cannot be manipulated through collusion, as each client autonomously embeds their identifier without relying on centralized watermark management that is vulnerable to coordinated attacks.
3Loss of information
If digital watermarking is implemented at the back end before encoding, then piracy source identification is enabled, but the identification granularity is limited to coarse levels such as content distribution services
Solution Approach 1:
The patent reverses the traditional approach by implementing watermarking after decoding on the client side rather than before encoding at the back end. This inversion enables the embedding of fine-grained client-specific identifiers that survive piracy attacks, allowing identification at the individual user account level rather than just at the content distribution service level.
Solution Approach 2:
The system performs preliminary watermarking on the decoded video frames before the video is rendered or potentially pirated. By embedding the unique client identifier at this intermediate stage in the playback pipeline, the watermark is preserved through subsequent encoding and distribution, enabling precise source identification while maintaining the ability to operate within DRM trust zones.
Data Source
AI summary
Techniques are described for associating identifiers (e.g., digital watermarks) with video content in a way that enables identification of the source of pirated content with specificity as granular as an individual user account. A compositors operating in the DRM trust zone of a client device introduces the identifier by compositing overlay information with decoded video frames. The identifier may then be recovered by comparing target content to the source content to extract the overlay information.