首页>
外国专利>
Secure audio watermarking based on neural networks
Secure audio watermarking based on neural networks
展开▼
机译:基于神经网络的安全音频水印
展开▼
页面导航
摘要
著录项
相似文献
摘要
Embodiments provide systems, methods, and computer storage media for secure audio watermarking and audio authentication. An audio watermark detector may include a neural network trained to detect a particular audio watermark and embedding technique, which may indicate source software that has been used in a workflow that generated an audio file to be examined. The watermark can, for example, indicate that an audio file has been generated using voice manipulation software, so that the detection of the watermark can indicate manipulated audio, such as deepfake audio and other attacked audio signals. In some embodiments, the audio watermark detector can be trained as part of a generative-adverse network to make the underlying audio watermark more resistant to neural network-based attacks. In general, the audio watermark detector can evaluate time domain samples from portions of an audio clip to be examined in order to detect the presence of the audio watermark and generate a classification for the audio clip.
展开▼