作者: Vadim Shchemelinin , Konstantin Simonchik
DOI: 10.1007/978-3-319-01931-4_18
关键词:
摘要: This paper examines the method of spoofing text-dependent voice verification systems based on most popular TTS approaches: Unit Selection and HMM. Research this shows possibility achieving a false acceptance error 98%-100% if duration database is sufficiently large. A distinctive feature that it can be fully automatical used in conjunction with speech recognition system.