Browse Prior Art Database

Improving recognition accuracy for spoken passwords Disclosure Number: IPCOM000014820D
Original Publication Date: 2001-Jan-01
Included in the Prior Art Database: 2003-Jun-20
Document File: 2 page(s) / 54K

Publishing Venue



User goes to registration system.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 74% of the total text.

Page 1 of 2

Improving recognition accuracy for spoken passwords

User goes to registration system.

User receives randomly assigned password.

User calls application.

When system requests password, user speaks it.


Password recognized?


System goes to bad password management process.

User enters application.

Disclosed is a method of assigning passwords for users to speak to gain entry to a speech-enabled application. The fundamental concept is to provide randomly selected passwords to users, drawing the passwords from a pool of passwords known to have good recognition accuracy.

Some speech-enabled applications require a user to provide a password to enter the application. In some cases, this entry to the application depends on correct recognition of the user's name followed by a spoken password or DTMF PIN. The DTMF PIN has the advantage of not relying on a probabilistic process such as speech recognition, but is a less natural way for a user to interact with a system. Relying on correct recognition of a user's name and a user-generated password puts the system at risk because proper names are among the most difficult types of tokens for a system to recognize, and allowing users to generate their own passwords opens the


[This page contains 6 pictures or other non-text objects]

Page 2 of 2

door for a user to unknowingly create an impossible- or difficult-to-recognize password.

Using the disclosed method, a system automatically assigns a password to a subscribed user, and...