Blockchain Video Authentication via Segmented Hash Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Public-facing organizations are vulnerable to deepfakes, as generative machine learning algorithms can create synthetic videos that are difficult to detect, leading to the spread of disinformation by using their original media as training data.
Innovation Solution
A video registration and authentication system using a blockchain to store and verify the authenticity of videos, where source videos are registered with a hash value and compared to target videos to identify any modifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If source videos are registered in a blockchain to protect against deepfakes, then video authenticity and security are improved, but device complexity and processing overhead increase
Solution Approach 1:
The patent extracts only the essential authentication data (hash values, timestamps, metadata) from the complete video files and stores these extracted elements in the blockchain. This allows the system to maintain video authenticity verification capability while minimizing the data burden on the blockchain, thereby reducing system complexity while preserving reliability.
Solution Approach 2:
The patent introduces an intermediary authentication server that acts as a mediator between the blockchain and end users. This server handles the complex tasks of hash generation, blockchain interactions, and video verification, shielding users from direct blockchain complexity while maintaining secure authentication. The intermediary absorbs the computational overhead and presents a simplified interface.
2Measurement precision
If the blockchain stores complete video data for authentication, then measurement precision of video authenticity is improved, but storage requirements and processing time increase
Solution Approach 1:
The system extracts only the critical authentication elements (hash values representing video content, timestamps, and essential metadata) from complete video files for blockchain storage. This extraction maintains authentication accuracy by preserving the cryptographic fingerprint of the video while dramatically reducing the data volume that needs to be processed and stored, thereby minimizing processing time delays.
Solution Approach 2:
The patent transforms the video authentication problem from comparing entire video files to comparing compact hash values and metadata parameters. By changing the parameter representation from full video data to cryptographic hashes, the system achieves the same authentication precision with significantly reduced processing time and storage requirements.
3Measurement precision
If hash values are generated for every video segment for blockchain storage, then detection precision of video modifications is improved, but computational resources and energy consumption increase
Solution Approach 1:
The patent segments videos into manageable portions and generates hash values for each segment rather than processing entire videos as single units. This segmentation enables precise detection of modifications in any specific segment while distributing the computational workload across multiple smaller, independent hash generation tasks, thereby reducing peak energy consumption and computational overhead.
Solution Approach 2:
The system applies partial action by generating hash values only for the necessary video segments that require authentication, rather than uniformly processing all video data. This selective approach maintains high detection precision for modified segments while minimizing unnecessary computational energy expenditure on unmodified or less critical portions of video content.
Data Source
AI summary
A system includes a registration server and an authentication server. The registration server receives a source video and splits it into a series of segments that includes a first segment. The registration server generates a first hash from the first segment and stores it in a block in a blockchain. The registration server stores an identifier for the block in the first set of metadata. The authentication server receives a request to authenticate a target video and splits it into a series of segments that includes a first segment, corresponding to the first segment of the source video. The authentication server generates a second hash from the first segment. The authentication server determines that the second set of metadata includes the identifier and uses it to locate the block in the blockchain. The authentication server determines that the second hash does not match the first hash and transmits an alert.