kumoh national institute of technology
Networked Systems Lab.

Muhammad Rasyid Redha Ansori, Allwinnaldo, Revin Naufal Alief, Ikechi Saviour Igboanusi, Jae Min Lee, Dong-Seong Kim, "HADES: Hash-based Audio Detection System for Copyright Protection in Decentralized Music Sharing", IEEE Transactions on Network and Service Management Journal (Major Revision)
By : Rasyid
Date : 2022-05-03
Views : 78

Preventive measures to stop copyright infringement are still not in place on the current decentralized music sharing platform. As there is no detection technique to detect modified audio before it goes online, some decentralized music platforms become a place full of pirated audio files. A perceptual hash-based audio detection for copyright protection in decentralized music sharing is proposed to address this problem. Chromaprint, an open-source audio fingerprint program, is used to generate a perceptual hash to detect copyright infringement. To assess the perceptual hash technique's robustness, Chromaprint generates perceptual hashes from five audio files, and each of the audio is attacked with 14 kinds of signal processing attacks. The detection system's results show that it is very good at spotting copyright infringement, with an average match rate of 94.621%. Tested with a private Ethereum blockchain, the total execution time from hashing to updating the information to the blockchain takes 2,3213 ms.