构建多语言歌唱深度伪造检测数据集，助力真实场景下模型评估。
SingFox: A Multi-Lingual Singfake Detection Corpus
- 构建6类不同挑战的多语言歌唱伪造数据集
- 涵盖20种语言超11万段音频，总时长超126小时
- 支持检测与溯源双重任务，适合安全与语音分析研究者
本文提出SingFox，一个大规模、多语言的歌唱深度伪造检测与来源追溯数据集。该数据集分为六个独立模块（T1–T6），涵盖语言多样性（全球及印度语）、流派特异性音乐以及多种伪造生成方式等不同挑战。数据集包含超过113,802段音频，覆盖20种语言，总时长达126.32小时，涉及1,150位歌手。各模块模拟真实应用场景，用于评估模型在不同条件下的鲁棒性。实验表明，在跨数据集评估中最高准确率达77.84%。所有代码与资源已公开于https://github.com/Arth-Shah/SingFox，旨在推动可复现研究并加速该领域发展。
原文摘要 · Abstract (English)
In this work, we introduce SingFox, a comprehensive and large-scale dataset specifically designed to support robust evaluation of singing deepfake detection and source tracing systems. SingFox is divided into six distinct tracks (T1--T6), each targeting a unique form of novelty, ranging from language diversity (global and Indian) to genre-specific music and alternative fake generation methods. The dataset encompasses over 113,802 audio clips across 20 languages, totaling more than 126.32 hours of audio data and featuring 1,150 singers. Each track is designed to emulate real-world scenarios and evaluate how reliably models perform under different conditions, thereby assessing their robustness. SingFox aims to foster reproducibility and accelerate research in singing deepfake detection by providing a reliable benchmark for both the singfake detection task and the source verification task (model explainability). Experimental results show a highest accuracy of 77.84\% in cross-dataset evaluation settings. All code and resources required to reproduce the dataset are publicly available at https://github.com/Arth-Shah/SingFox.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。