ASSERT: Anti-Spoofing with Squeeze-Excitation and Residual neTworks

Abstract

We present JHU's system submission to the ASVspoof 2019 Challenge: Anti-Spoofing with Squeeze-Excitation and Residual neTworks (ASSERT). Anti-spoofing has gathered more and more attention since the inauguration of the ASVspoof Challenges, and ASVspoof 2019 dedicates to address attacks from all three major types: text-to-speech, voice conversion, and replay. Built upon previous research work on Deep Neural Network (DNN), ASSERT is a pipeline for DNN-based approach to anti-spoofing. ASSERT has four components: feature engineering, DNN models, network optimization and system combination, where the DNN models are variants of squeeze-excitation and residual networks. We conducted an ablation study of the effectiveness of each component on the ASVspoof 2019 corpus, and experimental results showed that ASSERT obtained more than 93% and 17% relative improvements over the baseline systems in the two sub-challenges in ASVspooof 2019, ranking ASSERT one of the top performing systems. Code and pretrained models will be made publicly available.

Submitted to Interspeech 2019, Graz, Austria

References in corpus (1)

• The BOSARIS Toolkit: Theory, Algorithms and Code for Surviving the New DCF

Cited by in corpus (4)

• Adversarial Attacks on Spoofing Countermeasures of automatic speaker verification
• Integrated Replay Spoofing-aware Text-independent Speaker Verification
• Capturing scattered discriminative information using a deep architecture in acoustic scene classification
• Automatic Speech Verification Spoofing Detection