Recent advancements in automatic speaker authentication

摘要

Verbal-information and fixed-phrase approaches for identifying a person via voice are ready for real-world applications. We focus exclusively on the use of voice for authentication applications and review advancements in this area. The technical components in speech recognition and verification systems are reviewed, and we then discuss a speech-verification system that utilizes stochastic matching to identify a person based on voice characteristics. We also discuss a verbal-information verification system that verifies identity through the content of the verbal information.