Detecting Replay Attacks Using Single-Channel Audio: The Temporal Autocorrelation of Speech

In this paper, we propose to use the temporal auto-correlation of single-channel speech as a new feature for replay detection. Visual comparisons show that the proposed feature distinguishes replay attacks from clean speech and speech with simulated reverberation. Experimental results on the ASVspoof 2019 physical access database show that the proposed feature contains crucial information against replay attacks and that using the proposed feature in a fusion system almost always leads to performance improvements. Furthermore, our best fusion system achieves equal error rate and minimum tandem detection cost function of 0 on the development set for the first time.

Automated summary

This paper proposes a new feature for replay detection, which utilizes the temporal auto-correlation of single-channel speech. The experimental results demonstrate that the proposed feature can effectively distinguish replay attacks, clean speech, and speech with simulated reverberation, and its utilization in a fusion system consistently improves performance. Moreover, the best fusion system achieves a zero equal error rate and a zero minimum tandem detection cost function for the first time on the development set.