Solving Copyright Infringement on Short Video Platforms: Novel Datasets and an Audio Restoration Deep Learning Pipeline

2025-04-30

AI Summary

To address copyright evasion in short-video platforms—where infringers embed arbitrary background music (BGM) to mask original soundtracks (OSTs) and bypass detection—this paper proposes an end-to-end audio restoration framework. First, a short-horizon-adapted deep music source separation model robustly isolates interfering BGM; then, cross-modal audio-video retrieval combined with multi-granularity alignment modeling enables semantically verifiable OST recovery. We introduce OASD-20K and OSVAR-160, the first short-horizon-specific datasets designed for copyright compliance, and propose a joint optimization paradigm for BGM-robust separation and verifiable OST restoration. Evaluated on OSVAR-160, our method achieves 92.7% BGM removal accuracy and 89.4% OST semantic fidelity, significantly enhancing copyright traceability for user-generated content and improving the efficacy of automated platform-level content moderation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Adversarial Attacks & Robustness
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

Short video platforms like YouTube Shorts and TikTok face significant copyright compliance challenges, as infringers frequently embed arbitrary background music (BGM) to obscure original soundtracks (OST) and evade content originality detection. To tackle this issue, we propose a novel pipeline that integrates Music Source Separation (MSS) and cross-modal video-music retrieval (CMVMR). Our approach effectively separates arbitrary BGM from the original OST, enabling the restoration of authentic video audio tracks. To support this work, we introduce two domain-specific datasets: OASD-20K for audio separation and OSVAR-160 for pipeline evaluation. OASD-20K contains 20,000 audio clips featuring mixed BGM and OST pairs, while OSVAR160 is a unique benchmark dataset comprising 1,121 video and mixed-audio pairs, specifically designed for short video restoration tasks. Experimental results demonstrate that our pipeline not only removes arbitrary BGM with high accuracy but also restores OSTs, ensuring content integrity. This approach provides an ethical and scalable solution to copyright challenges in user-generated content on short video platforms.

Problem

• Detect and separate unauthorized background music in short videos
• Restore original soundtracks to ensure content authenticity
• Address copyright infringement challenges on platforms like TikTok

Innovation

• Music Source Separation for BGM removal
• Cross-modal video-music retrieval integration
• Domain-specific datasets for audio restoration