An open speaker recognition enabled identification and authentication system

The development and roll-out of ICT and information systems in Africa has steadily increased in recent years as African nations attempt to bridge the gap between the digital divide. This increasing reliance on information systems has introduced many security risks, the existence of which has prompted the greater necessity to protect these systems. The use of knowledge-based, token-based and biometrics-based authentication security controls have contributed greatly in protecting information assets. However, it is also known that these systems have become increasingly susceptible to be being compromised. This reality has inspired research into additional authentication methods. Speaker recognition is one such method that offers many unique benefits, however it is not widely used. There are also no open and complete speaker recognition systems that could be found that may assist in the research process. This paper will provide an overview of speaker recognition along with the potential benefits and limitations, many of which are applicable in the African context. This paper also present an open identification and authentication enabled speaker recognition software prototype intended to serve as a platform for speaker recognition learning, experimentation and research.