Automatic speaker verification systems are vulnerable to audio replay attacks which bypass security by replaying recordings of authorized speakers. Replay attack detection (RA) systems built upon Residual Neural Networks (ResNet)s have yielded astonishing results on the public benchmark ASVspoof 2019 Physical Access challenge. With most teams using fine-tuned feature extraction pipelines and model architectures, the generalizability of such systems remains questionable though. In this work, we analyse the effect of discriminative feature learning in a multi-task learning (MTL) setting can have on the generalizability and discriminability of RA detection systems. We use a popular ResNet architecture optimized by the cross-entropy criterion as our baseline and compare it to the same architecture optimized by MTL using Siamese Neural Networks (SNN). It can be shown that 26.8% relative improvement on Equal Error Rate (EER) is obtained by leveraging SNN.We further enhance the model’s architecture and demonstrate that SNN with additional reconstruction loss yield another significant improvement of relative 13.8% EER.
Cite as: Platen, P.v., Tao, F., Tur, G. (2020) Multi-Task Siamese Neural Network for Improving Replay Attack Detection. Proc. Interspeech 2020, 1076-1080, doi: 10.21437/Interspeech.2020-86
@inproceedings{platen20_interspeech,
author={Patrick von Platen and Fei Tao and Gokhan Tur},
title={{Multi-Task Siamese Neural Network for Improving Replay Attack Detection}},
year=2020,
booktitle={Proc. Interspeech 2020},
pages={1076--1080},
doi={10.21437/Interspeech.2020-86}
}