1d-cnn-based Audio Tampering Detection Using Enf Signals

This article discusses the use of electric network frequency (ENF) as a reliable technique in audio forensics and introduces a blind audio forensics framework, ATD, based on a one-dimensional convolutional neural network (1D-CNN) model. The framework incorporates characteristics of ENF signals and a denoising method to enhance feature extraction and reduce the impact of external noise. Extensive experiments confirm the effectiveness, efficiency, adaptability, and robustness of ATD in detecting audio tampering. The article also highlights the importance of security and privacy protection in the era of the Internet of Things (IoT).