Customer identification through voice biometrics

Patent metadata

Field	Value
Publication number	US-9607621-B2
Application number	US-201615180458-A
Country	US
Kind code	B2
Filing date	Jun 13, 2016
Priority date	Sep 30, 2013
Publication date	Mar 28, 2017
Grant date	Mar 28, 2017

Abstract

Systems and methods for determining an identity of an individual are provided. Audio may be received that includes a key phrase spoken by the individual, and the key phrase may include an identifier spoken by the individual. A key phrase voice print and key phrase text corresponding to the audio may be obtained. The key phrase text may include text corresponding to the identifier spoken by the individual. Voice prints may be retrieved based on the text corresponding to the identifier, and the voice prints may be provided to a voice biometric engine for comparison to the key phrase voice print. The individual may be authenticated based on a comparison of the key phrase voice print to the voice prints. The identifier may include a first name and a last name of the individual.

First claim

What is claimed is:

1. A non-transitory computer-readable media having instructions stored thereon that, when executed by a processor of a computing device, cause the computing device to perform steps for determining an identity of an individual, the steps comprising:
receiving, in a communication via a communication portal, audio that includes a key phrase spoken by the individual and the key phrase includes an identifier spoken by the individual;
obtaining a key phrase voice print corresponding to the audio;
obtaining key phrase text corresponding to the audio wherein the key phrase text includes text corresponding to the identifier spoken by the individual;
retrieving a set of voice prints based on the text corresponding to the identifier;
determining that a total number of voice prints in the set of voice prints exceeds a predetermined size threshold;
selecting at least one of the voice prints in the set of voice prints to exclude from comparison to the key phrase voice print; and
providing the key phrase voice print and one or more voice prints in the set of voice prints, other than the at least one voice print selected for exclusion, to a voice biometric engine for comparison.

2. The non-transitory computer-readable media of claim 1 , wherein:
the identifier comprises a first name and a last name of the individual; and
each of the one or more voice prints retrieved is respectively associated with a customer profile that includes a profile first name that matches the first name and a profile last name that matches the last name.

3. The non-transitory computer-readable media of claim 1 , wherein:
the key phrase voice print is obtained from a voice print engine that receives the audio;
the key phrase text is obtained from a speech-to-text engine that receives the audio; and
the one or more voice prints are retrieved from a voice print database that stores the one or more voice prints.

4. The non-transitory computer-readable media of claim 1 , wherein the communication portal is one of an interactive voice response portal, a mobile portal, and an online portal.

5. The non-transitory computer-readable media of claim 1 , wherein:
the at least one voice print selected for exclusion is associated with at least one previous communication; and
the at least one voice print selected for exclusion is selected responsive to a determination that a characteristic of the communication does not match a previous characteristic of the at least one previous communication.

6. The non-transitory computer-readable media of claim 5 , wherein the characteristic and the previous characteristic are one of:
a phone number associated with the communication and a previous phone number associated with the at least one previous communication;
a network address associated with the communication and a previous network address associated with the previous communication; and
a device identifier associated with the communication and a previous device identifier associated with the previous communication.

7. A system for determining an identity of an individual comprising:
one or more processors; and
memory storing computer readable instructions that, when executed by one of the processors, cause the system to:
receive, in a communication via a communication portal, audio comprising a key phrase spoken by an individual, the key phrase comprising an identifier spoken by the individual,
obtain a key phrase voice print corresponding to the audio,
convert the audio to text, the text comprising key phrase text corresponding to the key phrase and the key phrase text comprising identifier text corresponding to the identifier,
query a voice print database with the identifier text to obtain a set of voice prints associated with the identifier,
responsive to determining that a total number of voice prints in the set of voice prints exceeds a predetermined size threshold, select at least one of the voice prints in the set of voice prints to exclude from comparison to a key phrase voice print; and
provide the key phrase voice print and the set of voice prints, other than the at least one voice print selected for exclusion, to a voice biometric engine for comparison.

8. The system of claim 7 , wherein:
the identifier comprises a first name and a last name of the individual; and
each voice print of the set of voice prints is associated with a customer profile, the customer profile comprising a profile first name that matches the first name and a profile last name that matches the last name.

9. The system of claim 8 , wherein:
the instructions, when executed by one of the processors, further cause the system to obtain a variant spelling of the first name; and
the identifier text comprises first text corresponding to the first name and second text corresponding to the variant spelling of the first name.

10. The system of claim 8 , wherein:
the instructions, when executed by one of the processors, further cause the system to obtain a variant spelling of the last name; and
the identifier text comprises first text corresponding to the last name and second text corresponding to the variant spelling of the last name.

11. The system of claim 8 , wherein:
the instructions, when executed by one of the processors, further cause the system to obtain a first variant spelling of the first name and a second variant spelling of the last name; and
the identifier text comprises first text corresponding to the first name and the last name and comprises second text corresponding to the first variant spelling of the first name and the second variant spelling of the last name.

12. The system of claim 7 , wherein the instructions, when executed by one of the processors, further cause the system to:
compare, using the voice biometric engine, the key phrase voice print to each voice print in the set of voice prints; and
receive, from the voice biometric engine a set of confidence scores, each confidence score indicating the extent to which the key phrase voice print matches one of the voice prints.

13. The system of claim 12 , wherein the instructions, when executed by one of the processors, further cause the system to:
determine which confidence score in the set of confidence scores is the highest confidence score;
determining whether the highest confidence score is greater than an upper confidence threshold and whether the highest confidence score is lower than a lower confidence threshold; and
responsive to determining that the highest confidence score is greater than the upper confidence threshold, authenticating the individual.

14. The system of claim 13 , wherein the instructions, when executed by one of the processors, further cause the system to:
grant the individual access to one or more services provided by a banking system subsequent to authenticating the individual.

15. The system of claim 13 , wherein the instructions, when executed by one of the processors, further cause the system to:
responsive to determining that the highest confidence score is less than the upper confidence threshold and greater than the lower confidence threshold, prompt the individual to answer one or more security questions; and
responsive to determining the individual correctly answered a predetermined threshold number of the one or more security questions, authenticate the individual.

16. The system of claim 7 , wherein:
the voice print selected for exclusion is associated with at least one previous communication, and
the instructions, when executed by one of the processors, further cause the system to select

Assignees

• Bank Of America

Inventors

• Karpey David
• Pender Mark

Classifications

• G07C9/37

  using biometric data, e.g. fingerprints, iris scans or voice recognition · CPC title

• G10L17/24 Primary

  the user being prompted to utter a password or a predefined phrase · CPC title

• H04M3/493

  Interactive information services, e.g. directory enquiries {; Arrangements therefor, e.g. interactive voice response [IVR] systems or voice portals} · CPC title

• G10L15/26

  Speech to text systems (G10L15/08 takes precedence) · CPC title

• G10L17/26

  Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices · CPC title