| Speaker verification in WWW applications |
Title: Speaker verification in WWW applications
Author(s): Lou Boves & Johan Koolwaaij
Reference: Proceedings of the Workshop "Reconnaissance du Locuteur et ses Applications Commerciales et Criminalistiques" (RLA2C), Avignon, France, 20-23 April 1998, pp. 178-181
Keywords: Speaker Recognition
There is a PostScript version (172223 bytes) available.
There is a PDF version (107370 bytes) available.
Speaker Recognition (SR) has been studied for use in protecting physical access to premises, electronic access to data and information and to monitor the whereabouts of persons who are not allowed to travel freely. A comparison of these applications shows that each makes specific requirements to the technology and especially to the way in which the underlying recognition algorithms must be tuned and integrated in the overall hardware and software architecture.
Recently, interest in yet another application domain has grown. Increasingly, service providers are interested in using SR to protect access to Internet and World-Wide-Web services. In this paper we investigate the ways in which Web-based applications make specific requirements that have an impact on the implementation and the performance of the SR technology.
The paper is based on our experiences with access to a Web site created by
the Dept. of Language & Speech. This site does not contain confidential
information, and visitors do not receive special rewards when they access the
site, as true customers or as imposters. Thus, this paper does not aim
to describe a formal test of the level of security that can be attained
with the help of specific SR technology.
Error processing SSI file