Voiceprint identification system

• US 6,356,868 B1
• Filed: 10/25/1999
• Issued: 03/12/2002
• Est. Priority Date: 10/25/1999
• Status: Expired due to Term

• Current Assignee: Mavenir Systems Incorporated (Mavenir Group Holdings Ltd.)
• Original Assignee: Comverse Network Systems, Inc. (Verint Systems Incorporated)
• Inventors: Yuschik, Matthew J., Slezak, Robert J.
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Nolan, Daniel A.
• Application Number:
  US09/422,851
• Time in Patent Office: 869 Days
• Field of Search: 704/273, 704/231, 704/246, 704/251
• US Class Current: 704/246
• CPC Class Codes: G10L 17/14 Use of phonemic categorisation or speech recognition prior to speaker recognition or verification

Abstract

A voiceprint identification system identifies and verifies a user from voice data collected from a single interaction with the user. The voice data is a number, word phrase or any utterance chosen by the user. A first speech-processor processes the voice data to produce first match criteria, and a second speech-processor processes the same voice data to produce second match criteria, the first match criteria being different than the second match criteria. The first match criteria is used to select a subset of authorized persons, and, for each selected authorized person, the authorized person'"'"'s voice template is retrieved from a database. The retrieved voice templates are individually compared to the second match criteria until either the second match criteria matches one of the retrieved voice templates or all the retrieved voice templates have been compared without matching the second match criteria.

Claims

1. A voiceprint identification system, comprising:

a speaker-independent processor receiving and converting a spoken passcode chosen by a user into a corresponding passcode identifier;

a speaker-dependent processor receiving and converting the spoken passcode to a voice template; and

said speaker-independent and speaker-dependent processors operating simultaneously.

27. A voiceprint identification system as recited in claim 1, wherein the spoken passcode is chosen by the user.

2. A voiceprint identification system, comprising:

a database storing a record for each user including a speech template and a passcode identifier;

a first processor receiving and converting a spoken passcode chosen by a user into a corresponding passcode identifier; and

a second processor comparing said spoken passcode to speech templates stored in a corresponding subset.

3. A voiceprint identification system as recited in claim 2, wherein the spoken passcode chosen by the user can be any audio signal, including any sound created through variations of the articulatory mechanism, including a word or non-word in any language.

4. A voiceprint identification system as recited in claim 2, wherein said record includes a telephone number;

and said second processor first compares the record in the subset with the phone number that matches a users calling number.

5. The voiceprint identification system as claimed in claim 2, wherein said record includes a phone number;

and said second processor first compares the records in the subset with an area code or country code of a users calling number.

6. The voiceprint identification system as claimed in claim 2, wherein said database stores error threshold settings;

said record includes a phone number; and

said second processor adjusts said error threshold settings to allow greater margin for error when a users calling number is identical to the phone number stored in the record being compared.

7. A voiceprint identification system as recited in claim 2, further comprising a training apparatus which activates each time the user speaks the spoken passcode to gain access.

8. A voiceprint identification system as recited in claim 2, wherein said record includes error threshold settings;

and said second processor adjusts said error threshold settings for each user in the subset when a new user is initialized.

9. A voiceprint identification system as recited in claim 2, wherein the second processor stops searching the subset once a comparison results in a match.

10. A voiceprint identification system as recited in claim 2, wherein the second processor completes searching the entire subset and returns a best match.

11. A voiceprint identification system as recited in claim 2, wherein multiple users can share the same account.

12. The voiceprint identification system as claimed in claim 2, wherein upon one or more unsuccessful attempts by the user to access an account corresponding to said passcode identifier, the system will store user'"'"'s calling number and spoken passcode.

13. A voiceprint identification database system, comprising:

passcode identifier data stored on a computer readable medium;

biometric data associated with its respective passcode identifier stored on a computer readable medium; and

a processor reading and writing the passcode identifier data and the biometric data from and to the computer readable medium,

wherein the passcode identifier data and the biometric data are indexed on the computer readable storage medium according to the passcode identifier data, so that a spoken passcode is used for indexing a set of passcode identifier data by converting the spoken passcode into the passcode identifier data.

14. A voiceprint identification database as recited in claim 13, wherein said database stores additional information particular to an individual along with said biometric data.

15. A voiceprint identification system, comprising:

a database storing a record for each user, the record including a speech template, a passcode identifier, a telephone number, biometric data and error threshold information;

a first processor receiving and converting the spoken passcode into its corresponding passcode identifier;

a second processor comparing said spoken passcode to speech templates stored in a corresponding subset;

said spoken passcode chosen by the user can be any audio signal in addition to a spoken number or a word phrase;

said second processor first compares the record in the subset with the phone number that matches the users calling number;

said second processor compares the records in the subset with an area code identical to an area code of the users calling number;

said second processor adjusts error threshold settings to allow greater margin for error when users calling number is identical to phone number stored in record being compared;

training apparatus activating each time user speaks his spoken passcode to gain access;

said second processor adjusting error threshold settings for each user in the subset when a new user is initialized in order to assure discrimination between records in the subset;

said second processor can be set to stop searching the subset once a comparison results in a match;

said second processor can be set to complete searching the entire subset and return a best match;

a security device storing, upon one or more unsuccessful attempts by the user to access an account corresponding to said passcode identifier, the user'"'"'s calling number and spoken passcode; and

said passcode may be any sound created through variations of the articulatory mechanism, and is a word or non-word in any language.

16. A method of operating a voiceprint identification system, comprising:

allowing a plurality of users to choose their own respective spoken passcode;

storing records for each user, including a speech template and a passcode identifier;

receiving a spoken passcode by a user;

converting the spoken passcode to a passcode identifier; and

searching a subset of the passcode identifier for a match between the spoken passcode and the speech templates.

17. A computer readable storage media for controlling a computer to perform:

storing a voiceprint identification system process;

allowing a plurality of users to choose their own respective spoken passcodes;

storing records for each user including a speech template and a passcode identifier;

receiving a spoken passcode by a user;

converting the spoken passcode to a passcode identifier; and

searching a subset of the passcode identifier for a match between the spoken passcode and the speech templates.

18. An access control process, comprising:

converting a user defined spoken passcode into a digital passcode;

converting the spoken passcode into a spoken voice template;

retrieving a stored voice template using the digital passcode;

comparing the stored and spoken voice templates; and

indicating access is allowed when the stored and spoken voice templates match within a tolerance.

19. A process as recited in claim 18, wherein producing the digital passcode uses a speaker independent voice process and producing the voice template uses a speaker dependent voice process.

20. A voiceprint identification system for ascertaining whether a user is an authorized person by processing speech from the user, the voiceprint identification system comprising:

a first speech-processor to receive the speech and providing first match criteria;

a database for storing voice templates of authorized persons, the database being coupled to the first speech-processor such that the first match criteria select a subset of one or more of the stored voice templates;

a second speech-processor to receive a same speech as the first speech-processor and providing second match criteria, the first match criteria being different from the second match criteria; and

a comparator coupled to the database to receive the subset of voice templates and coupled to the second speech-processor to receive the second match criteria, the comparator providing a match indicator when the second match criteria matches exactly one of the subset of voice templates, otherwise the comparator providing a no-match indicator.

21. A method comprising:

converting an audible passcode into a digital passcode;

selecting a subset of records using the digital passcode; and

retrieving information related to the audible passcode from the selected subset.

22. A method as recited in claim 21, wherein the audible passcode can be be chosen by a user of the audible passcode.

23. A computer readable storage medium, storing a program instructing a computer to perform:

converting an audible passcode into a digital passcode;

selecting a subset of records using the digital passcode; and

retrieving information related to the audible passcode from the selected subset.

24. A computer readable storage medium as recited in claim 23, wherein the audible passcode can be chosen by a user of the audible passcode.

25. An apparatus comprising:

a converter converting an audible passcode into a digital passcode;

a selector selecting a subset of records using the digital passcode received from the converter; and

a retriever retrieving information related to the audible passcode from the subset selected by the selector.

26. An apparatus as recited in claim 25, further comprising a choosing unit allowing a user to select the user'"'"'s own audible passcode.