Edge Watermarking for Live Streaming Collusion Resistance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing A/B forensic watermarking schemes in VOD streaming are vulnerable to collusion attacks and require high computational resources for real-time user information encoding, leading to potential transmission delays and increased system costs.
Innovation Solution
A method that inserts user-specific watermarks (W watermarks) into segments at cyclic Nth positions using edge computing, allowing real-time forensic watermarking with resistance to collusion attacks, even with low-performance edge devices, by adjusting the value of N to manage computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If user information is inserted into every video frame using encoding processes, then forensic watermarking capability is improved, but computational load increases and transmission delay occurs
Solution Approach 1:
The patent divides the video stream into segments and applies watermarking only to specific segments (Nth segments) rather than all frames. This segmentation approach maintains forensic capability while reducing computational load and transmission delay by processing only a subset of the content.
Solution Approach 2:
The patent implements periodic watermarking where user information is inserted into video segments at cyclic Nth positions rather than continuously. This periodic action maintains sufficient forensic identification capability while significantly reducing computational resources and transmission time compared to continuous watermarking.
2Productivity
If A/B watermarking scheme is used in VOD streaming, then efficiency for simultaneous user access is improved, but vulnerability to collusion attacks increases
Solution Approach 1:
The patent pre-generates A and B watermark versions of video segments during the preprocessing stage, storing them in a database. When a user requests content, the system retrieves the appropriate pre-generated segments and combines them according to the A/B scheme, enabling efficient simultaneous access while maintaining security through proper segment selection and combination.
Solution Approach 2:
The patent applies different watermarking strategies to different segments - using A or B watermarks based on the specific segment position and user identification. This local differentiation allows efficient multi-user service while maintaining security by ensuring that colluding users cannot combine segments to destroy the watermark, as each user receives customized segment combinations.
3Reliability
If real-time watermarking is implemented for all users, then forensic identification is improved, but system cost and computational resources increase
Solution Approach 1:
The patent applies partial watermarking by inserting user information only into Nth segments rather than all segments. This partial action provides sufficient forensic identification capability for practical purposes while significantly reducing computational resource consumption and system complexity compared to complete real-time watermarking of all content.
Solution Approach 2:
The system performs preliminary watermark generation for A and B versions during preprocessing, storing the results in a database. During actual streaming, the system only needs to retrieve and combine pre-generated segments rather than performing complex real-time encoding for all users, dramatically reducing computational resource requirements while maintaining forensic identification capability.
Data Source
AI summary
The present invention provides a system and a method for providing real-time forensic watermarking using edge computing in a live streaming environment. The pre-processing device receives original segments of multimedia content, inserts an A watermark and a B watermark according to the A/B watermarking scheme into each original segment, and generate an A content and a B content. Herein, every segment at the cyclic Nth positions of both the A content and the B content is retained as the original segment. The edge device generates a W watermark using predetermined session information, inserts the W watermark into the original segment at the cyclic Nth positions, and encodes it into a W segment. The CDN device transmits the A segment or the B segment according to the order determined by the A/B watermarking scheme for the 1st to (N−1)th segments, and transmits the W segment for the Nth segment.