.

Jinzi Li1, Hexu Wang1,2, Fei Xie3,4

  • 1Xi'an Key Laboratory of Human-Machine Integration and Control Technology for Intelligent Rehabilitation, Xijing University, Xi'an 710123, China.

PubMed
概括

这项研究介绍了一种新的双路径时间频率注意网络 (DPTFAN),用于打击音频深度伪造. 该方法提高了检测准确度和对噪声和压缩的稳定性,提高了信息安全性.