An Investigation about the Scalability of the Spoofing Detection System
Gajan Suthokumar, Kaavya Sriskandaraja, Vidhyasaharan Sethu, Chamith Wijenayake, Eliathamby Ambikairajah · 2018
Anti spoofing techniques to mitigate spoofing threats in automatic speaker verification has gained more interest. Spoofing detection system is supposed to work for any kind of spoofing attacks in the real world. Nowadays, Spoofing countermeasures built either to mitigate synthetic speech or replayed speech but not for both simultaneously. This work investigates about the effectiveness of building a unified spoofing detection system which could handle the synthesized speech and replayed speech. This is first ever attempt in this research field. We have used the state of the art system features and classifiers for this study. All the experiments are done on ASVSpoof2015 and ASVSpoof 2017 V2.0 standard datasets.