BioLip: 通过生物力学约束违反建模实现语言泛化的唇同步深度伪造检测
BioLip: Language-Generalizable Lip-Sync Deepfake Detection via Biomechanical Constraint Violation Modeling
AI总结：
针对现有检测方法在生成器或语言迁移下失效的问题，提出基于唇部运动生物力学约束的轻量级三分支网络，仅利用地标坐标检测唇同步伪造，在零样本设置下对未见生成器和多种语言表现鲁棒。
摘要：
现有的唇同步深度伪造检测器依赖于像素伪影或视听对应关系，两者在生成器或语言迁移下均会失效，因为它们学习的特征与训练分布绑定。我们采用不同的方法。真实的唇部运动受到组织力学和神经肌肉带宽的约束；当前的生成器通常不施加这些约束，产生的轨迹在速度、加速度和加加速度上具有升高的方差，而真实语音不会表现出这些特征。我们利用这一信号，称为时间唇部抖动，通过从64个口周地标在短滑动窗口上计算运动学统计量，并将其输入一个轻量级三分支网络。该模型仅使用地标坐标：无像素、无音频、无声纹数据。我们仅在英语数据上训练，并在零样本设置下对五个未见生成器和七种语言进行测试。
英文摘要：
Existing lip-sync deepfake detectors rely on pixel artifacts or audio-visual correspondence, and both fail under generator or language shift because the features they learn are tied to the training distribution. We take a different approach. Authentic lip motion is constrained by tissue mechanics and neuromuscular bandwidth; current generators typically do not impose these constraints, producing trajectories with elevated variance in velocity, acceleration, and jerk that real speech does not exhibit. We exploit this signal, which we term temporal lip jitter, by computing kinematic statistics from 64 perioral landmarks over short sliding windows and feeding them into a lightweight three-branch network. The model uses only landmark coordinates: no pixels, no audio, and no voiceprint data. We train only on English data and test in a zero-shot setting on five unseen generators and seven languages.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
