Browse Prior Art Database

Original Publication Date: 2000-Mar-01
Included in the Prior Art Database: 2003-Jun-19

Publishing Venue



Disclosed is the use of text-independent speaker recognition and continuous speech recognition (at least in command and control form) to build a secure and user friendly Voice Cybervault for logins and password to multiple local and remote services. Recently with multiple machines, softwares and remote services offered on the internet and intranet, it is becoming extremely difficult for an average user to keep track of all the user Ids, logins and passwords or password procedures associated with each service. Besides being frustrating to the user, it leads to serious security breaches. Indeed, in order to remember the information appropriately, the user will either always use the same password for all the services or write down or store the information. In the first case if one login is compromised most of the other services accessible to the user becomes also compromised. In the second case, if the password booklet, paper or file is lost, all the services are compromised. Different alternatives have been proposed in the literature or on the market place to solve these problems. Biometrics are a first approach which allows to avoid passwords and even logins provided that the performances of the system are good enough. With the current state of the art, biometric-based access to a large network of numerous users is still shaky and needs to be complemented by password or knowledge-based information.