Method and system for bio-metric voice print authentication

• US 7,536,304 B2
• Filed: 05/24/2006
• Issued: 05/19/2009
• Est. Priority Date: 05/27/2005
• Status: Active Grant

• Current Assignee: Porticus Technology, Inc.
• Original Assignee: Porticus Incorporated
• Inventors: Salna, Bernardas, Di Mambro, Germano
• Primary Examiner(s): Vo; Huyen X.
• Application Number:
  US11/420,190
• Publication Number:
  US 20070185718A1
• Time in Patent Office: 1,091 Days
• Field of Search: 704/231, 704/235, 704/236, 704/243, 704/246, 704/273, 704/270, 704/270.1, 704/275, 704/250, 704/255, 704/205, 704/207, 704/209, 455/411, 455/410, 707/9
• US Class Current: 704/273
• CPC Class Codes: G06F 21/32 using biometric data, e.g. fingerprints, iris scans or voiceprints
• CPC Class Codes: G07C 9/37 using biometric data, e.g. fingerprints, iris scans or voice recognition
• CPC Class Codes: G10L 17/10 Multimodal systems, i.e. based on the integration of multiple recognition engines or fusion of expert systems

Abstract

A method (700) and system (900) for authenticating a user is provided. The method can include receiving one or more spoken utterances from a user (702), recognizing a phrase corresponding to one or more spoken utterances (704), identifying a biometric voice print of the user from one or more spoken utterances of the phrase (706), determining a device identifier associated with the device (708), and authenticating the user based on the phrase, the biometric voice print, and the device identifier (710). A location of the handset or the user can be employed as criteria for granting access to one or more resources (712).

Claims

1. A method for voice authentication on a device, comprising:

receiving one or more spoken utterances from a user at the device;

recognizing a phrase corresponding to said one or more spoken utterances;

identifying a biometric voice print from the one or more spoken utterances of said phrase;

verifying a location of said device;

determining a device identifier associated with said device;

authenticating said user based on the phrase, said biometric voice print, said location, and said device identifier;

establishing a connection to at least one authentication server;

sending a user profile to said at least one authentication server;

comparing said user profile with a plurality of reference profiles stored on said at least one authentication server;

upon recognizing the phrase, evaluating one or more vocal tract configuration differences between said spoken utterances;

matching one or more vocal tract shades of the plurality of reference profiles based on said vocal tract configuration difference;

determining at least one formant frequency;

determining at least one anti-formant frequency;

evaluating a difference between the formant frequency and the anti-formant frequency for characterizing one aspect of a natural change in a vocal tract configuration; and

determining if the user profile matches one of the plurality of reference profiles for authenticating said user,

wherein a user profile includes at least one of a phrase, a biometric voice print, and a device identifier.

2. The method of claim 1, wherein the identifying further comprises:

determining a variability of the one or more spoken utterances for creating said biometric voice print, wherein an acceptable variability of the biometric voice print is determined between at least two or more spoken utterances,

wherein the biometric voice print is a vocal tract configuration that is physically unique to a vocal tract of said user wherein the biometric voice print captures physical features associated with characteristic changes of a person'"'"'s vocal tract that are unique to an individual during the pronunciation of a pass phrase.

3. The method of claim 1, further comprising:

upon authenticating said user, granting access to one or more resources having a communication with the device.

4. The method of claim 1, wherein the step of identifying the biometric voice print further comprises the steps of:

generating a first voice print in response to a speaker'"'"'s voice;

generating at least a second voice print in response to the speaker'"'"'s voice;

identifying a difference between the first voice print and the second voice print;

determining whether the difference corresponds to a natural change of the speaker'"'"'s vocal tract; and

authenticating the speaker if the difference is indicative of a natural change in the speaker'"'"'s vocal tract.

5. The method of claim 1, wherein the device includes a global positioning system (GPS) for identifying the location of the device.

6. The method of claim 1, further comprising:

receiving a second spoken utterance by the user that identifies a location of the user; and

acknowledging that the location of the user corresponds to an accepted location.

7. The method of claim 1, wherein the mobile device identifier is an International Mobile Equipment Identity (IMEI) number.

8. The method of claim 7, wherein the IMEI number is associated with the biometric voiceprint in a database that uses the IMEI number for indexing the biometric voice print.

9. The method of claim 1, wherein the mobile device identifier is a PIN assigned to the device.

10. The method of claim 1, wherein the device is one of a mobile handset, a cell phone, a personal digital assistant, a laptop, a notebook, a portable music player, and a communications device.

11. The method of claim 1, wherein a first part of the authentication occurs on the device and a second part of the authentication occurs on the authentication server.

12. The method of claim 1, wherein the authentication server includes software applications such as a web layer, a business layer, and a database access layer.

13. The method of claim 1, wherein the vocal tract configuration differences are evaluated at one or more voiced segments of said spoken utterance.

14. The method of claim 1, wherein a vocal tract configuration difference corresponds to a bounded physical change of a user'"'"'s vocal tract associated with more than one spoken utterance.

15. The method of claim 1, wherein said vocal tract configuration difference is based on a momentum spectrum that accounts for a dynamic change of a speech spectrum over time.

16. The method of claim 15, wherein said momentum spectrum includes a lower bound and an upper bound for the one or more voice segments of speech, such that variations in the speech spectrum between the lower bound and the upper bound correspond to a unique vocal tract configuration.

17. The method of claim 1, wherein the vocal tract configuration is represented as one or more sections having a corresponding length and area that are characteristic to one or more sections of the user'"'"'s vocal tract.

18. The method of claim 1, further comprising:

determining whether differences between repeated spoken utterances are within a distortion range; and

determining a voice print match if the differences are within the distortion range.

19. The method of claim 1, wherein the evaluating a difference between the formant frequencies and anti-formant frequencies compensates for a telephone bandwidth.

20. The method of claim 1, wherein higher formant frequencies are emphasized to characterize one aspect of a speaker'"'"'s articulation style.

21. A method for voice authentication, comprising:

receiving a spoken utterance during a call on an Interactive Voice Response (IVR) system from a device;

invoking a programmatic action in response to a recognition of said spoken utterance,

analyzing said spoken utterance to identify a voice print of a user submitting said spoken utterance;

verifying a location of the device;

determining a device identifier associated with the device;

granting the user access to said programmatic action based on an authentication of a recognized voice print, a recognized spoken utterance, the device identifier, and the location of the device;

establishing a connection to at least one authentication server;

sending a user profile to said at least one authentication server;

comparing said user profile with a plurality of reference profiles stored on said at least one authentication server;

upon recognizing the phrase, evaluating one or more vocal tract configuration differences between said spoken utterances;

matching one or more vocal tract shades of the plurality of reference profiles based on said vocal tract configuration difference;

determining at least one formant frequency;

determining at least one anti-formant frequency;

evaluating a difference between the formant frequency and the anti-formant frequency for characterizing one aspect of a natural change in a vocal tract configuration; and

determining if the user profile matches one of the plurality of reference profiles for authenticating said user,

wherein a user profile includes at least one of a phrase, a biometric voice print, and a device identifier.

22. The method of claim 21, further comprising:

recognizing one or more numbers in the spoken utterance; and

automatically entering said one or more numbers for accessing a feature of the IVR,

whereby the programmatic action dials the one or more numbers recognized in the spoken utterance for accessing the feature.

23. The method of claim 21, further comprising:

from a device used to submit the spoken utterance, receiving a device identifier; and

validating a registration of the device to the user prior to granting access.

24. The method of claim 21, wherein the programmatic action is a secure transaction and the spoken utterance identifies secure information such as an account number, a bank routing number, a telephone number, a social security number, and a pin number.

25. The method of claim 24, further comprising:

upon said granting access, forwarding the call to a call center in accordance with the programmatic action; and

upon connection with the call center, invoking a screen pop at the call center to display user account information.

26. The method of claim 25, wherein the Interactive Voice Response (IVR) system resides on one of a PBX, a gateway, a voice server, and an internet voice server.