Blockchain assisted near-duplicated content detection
Résumé
This paper presents a study on the possibility of coupling blockchain solutions to multimedia tracking applications. The challenge lies in accommodating complex operations such as visual fingerprint extraction and management, which usually occur on general-purpose computing machines, under the blockchain framework. The advanced solution features a load- balancing architectural framework combining a multimedia app, its database, a Smart Contract, and a Token Contract. Thus, we bring forward the proof of concept for the tokenization of multimedia assets using this architecture. We also provide a resource utilization analysis for two use cases involving a robust video fingerprinting method and the International Standard Content Code, respectively. Thus, we demonstrate the mutually beneficial association of offchain and onchain applications for visual content tracking.