Joint Speaker Verification and Antispoofing in the i-Vector Space

Any biometric recognizer is vulnerable to spoofing attacks and hence voice biometric, also called automatic speaker verification (ASV), is no exception; replay, synthesis and conversion attacks all provoke false acceptances unless countermeasures are used. We focus on voice conversion (VC) attacks considered as one of the most challenging for modern recognition systems. To detect spoofing, most existing countermeasures assume explicit or implicit knowledge of a particular VC system and focus on designing discriminative features. In this work, we explore backend generative models for more generalized countermeasures. Specifically, we model synthesis-channel subspace to perform speaker verification and anti-spoofing jointly in the i-vector space, which is a well-established technique for speaker modeling. It enables us to integrate speaker verification and anti-spoofing tasks into one system without any fusion techniques. To validate the proposed approach, we study vocoder-matched and vocodermismatched ASV and VC spoofing detection on the NIST 2006 speaker recognition evaluation dataset. Promising results are obtained for standalone countermeasures as well as their combination with ASV systems using score fusion and joint approach.