Year: 2024
Type: conference-paper
Abstract: Music infringement detection has become a critical issue in the music industry, particularly in the digital area.With the rapid development of technology, the ways in which music is disseminated have drastically transformed.The rise of streaming platforms has made music easily accessible, but it has also provided fertile ground for infringement.Initially, infringement was limited to simple copying and pasting.However, it has since evolved into more covert and complex forms, including random clipping, secondary creation, and covers, which greatly complicates the challenge of copyright protection.To address this, the paper delves deeply into methods for detecting music infringement, aiming to safeguard music copyrights through technical means.For detecting direct infringement of original compositions, we combine music content classification with time-series matching techniques.This hybrid strategy effectively identifies unauthorized copying or usage by precisely capturing matching patterns.In the realm of cover music infringement detection, we employ deep learning methods based on music classification to obtain audio fingerprint features, which allow for in-depth analysis of the inherent characteristics of the music.The technique accurately assesses whether cover versions infringe upon the original creator's copyright.These methods not only improve detection accuracy but also enhance reliability, breathing new life into the field of music copyright protection.Experimental results demonstrate that the proposed approaches exhibit strong competitiveness across datasets, achieving high levels of accuracy and reliability.
Authors: Yang Zheng, Jie Liu, Shuwu Zhang
Language: English
Topic: Music and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4410858321