用5秒语音生成高保真声纹混变，无需重训练
VoxMorph: Scalable Zero-shot Voice Identity Morphing via Disentangled Embeddings
- 分离语调与音色嵌入，实现风格与身份精细控制
- 音频质量提升2.6倍，识别错误率降低73%
- 适合研究生物特征安全与语音伪造防御的读者
声纹混变技术可生成融合多人特征的伪造生物特征样本，使每位贡献者能通过单一注册模板被验证。尽管人脸混变已广泛研究，语音生物识别中的这一漏洞仍基本未被探索。现有语音混变方法计算成本高、难以扩展，且仅适用于声学特征相似的身份对，限制了实际应用。此外，现有声音混变方法针对音频纹理、音乐或环境声，无法迁移至声纹操控。本文提出VoxMorph，一种零样本框架，仅需每名参与者5秒音频即可生成高保真语音混变，无需模型重训练。该方法将语音特征解耦为语调与音色嵌入，通过球面线性插值（Slerp）融合，并使用自回归语言模型与条件流匹配网络合成语音。VoxMorph达到领先性能：音频质量提升2.6倍，可懂度错误率下降73%，在严格安全阈值下对自动说话人验证系统的攻击成功率达67.8%。本工作建立了实用且可扩展的语音混变范式，对生物特征安全具有重要意义。代码与数据集见项目页：https://vcbsl.github.io/VoxMorph/
原文摘要 · Abstract (English)
Morphing techniques generate artificial biometric samples that combine features from multiple individuals, allowing each contributor to be verified against a single enrolled template. While extensively studied in face recognition, this vulnerability remains largely unexplored in voice biometrics. Prior work on voice morphing is computationally expensive, non-scalable, and limited to acoustically similar identity pairs, constraining practical deployment. Moreover, existing sound-morphing methods target audio textures, music, or environmental sounds and are not transferable to voice identity manipulation. We propose VoxMorph, a zero-shot framework that produces high-fidelity voice morphs from as little as five seconds of audio per subject without model retraining. Our method disentangles vocal traits into prosody and timbre embeddings, enabling fine-grained interpolation of speaking style and identity. These embeddings are fused via Spherical Linear Interpolation (Slerp) and synthesized using an autoregressive language model coupled with a Conditional Flow Matching network. VoxMorph achieves state-of-the-art performance, delivering a 2.6x gain in audio quality, a 73% reduction in intelligibility errors, and a 67.8% morphing attack success rate on automated speaker verification systems under strict security thresholds. This work establishes a practical and scalable paradigm for voice morphing with significant implications for biometric security. The code and dataset are available on our project page: https://vcbsl.github.io/VoxMorph/
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。