Building whitelists comprising voiceprints not associated with fraud and screening calls using a combination of a whitelist and blacklist

• US 8,924,285 B2
• Filed: 01/12/2009
• Issued: 12/30/2014
• Est. Priority Date: 04/21/2005
• Status: Active Grant

• Current Assignee: Verint Americas Incorporated (Verint Systems Incorporated)
• Original Assignee: Verint Americas Incorporated (Verint Systems Incorporated)
• Inventors: Rajakumar, Anthony, Gutierrez, Richard, Guerra, Lisa M.
• Primary Examiner(s): Colbert, Ella
• Assistant Examiner(s): CRANFORD, MICHAEL D
• Application Number:
  US12/352,530
• Publication Number:
  US 20090119106A1
• Time in Patent Office: 2,178 Days
• Field of Search: 705/38
• US Class Current: 705/38
• CPC Class Codes: G06F 21/32 using biometric data, e.g. fingerprints, iris scans or voiceprints
• CPC Class Codes: G06F 21/577 Assessing vulnerabilities and evaluating computer system security
• CPC Class Codes: G06Q 20/24 Credit schemes, i.e. "pay after"
• CPC Class Codes: G06Q 20/341 Active cards, i.e. cards including their own processing means, e.g. including an IC or chip
• CPC Class Codes: G06Q 20/40 Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
• CPC Class Codes: G06Q 20/4014 Identity check for transactions
• CPC Class Codes: G06Q 20/40145 Biometric identity checks
• CPC Class Codes: G06Q 20/4016 involving fraud or risk level assessment in transaction processing
• CPC Class Codes: G07F 7/1008 Active credit-cards provided with means to personalise their use, e.g. with PIN-introduction/comparison system

Abstract

According to one aspect of the invention there is provided a method, comprising collecting voiceprints of callers; identifying which of the collected voiceprints are associated with fraud; and generating a whitelist comprising voiceprints corresponding to the collected voiceprints not identified as associated with fraud.

Claims

1. A method comprising:

receiving precursor data comprising a plurality of identifiers and voice samples, each identifier associated with a voice sample, the precursor data received without regard to whether the voice samples are fraudulent or not, the receiving performed using a server system including a computer having a processor coupled to memory;

receiving a fraud report comprising a plurality of identifiers associated with fraud;

identifying using the server system which of the precursor data are associated with fraud based on a match between identifiers associated with the voice samples and the identifiers associated with fraud in the fraud report;

generating a blacklist comprising a list of voice prints based on the voice samples corresponding to the identified precursor data associated with fraud using the server system;

generating a whitelist comprising a list of voice prints based on voice samples not corresponding to the identified precursor data associated with fraud using the server system; and

processing requests to screen voice samples of speakers against voice prints in one of the blacklist and the whitelist.

2. The method of claim 1, further comprising indexing the received precursor data based on identification metadata that at least partially identifies a speaker associated with the data.

3. The method of claim 1, wherein the generating a whitelist comprises converting precursor data not identified as associated with fraud to a whitelist database of voiceprints.

4. The method of claim 3, wherein the conversion is performed only after effluxion of a predefined period of time, provided that no fraudulent transaction was linked to the precursor data during the predefined period.

5. The method of claim 1, wherein generating a whitelist comprising data corresponding to the precursor data not identified as associated with fraud is based on a number of non-fraudulent transactions linked to the precursor data exceeds a threshold.

6. A system comprising:

a processor, and a memory coupled to the processor, the memory storing instructions which when executed by the processor cause the system to perform a method comprising;

receiving precursor data, each entry in the precursor data comprising an identifier and a voice print associated with the identifier, the data received without regard to whether the voice print is associated with fraud;

identifying which of the voice print entries in the precursor data are associated with fraud, based on matching the associated identifiers to information in a fraud report;

generating a blacklist comprising the voice print entries in the precursor data associated with fraud;

generating a whitelist comprising data corresponding to identifiers and associated voice print entries in the precursor data not identified as associated with fraud and based on identifiers associated with non-fraudulent transactions; and

processing requests to screen speakers against one of the blacklist and the whitelist.

7. The system of claim 6, wherein the method further comprises indexing the precursor data based on identification metadata that at least partially identifies a speaker associated with the data.

8. The system of claim 6, wherein the identifying comprises marking data associated with fraud if a fraudulent transaction can be linked to the data based on the fraud report.

9. The system of claim 6, wherein the generating comprises determining that a number of non-fraudulent transactions linked to the precursor data exceeds a threshold.

10. The system of claim 6, wherein the method further comprises generating a blacklist comprising voiceprints corresponding to precursor data identified as associated with fraud.

11. A non-transitory computer-readable storage medium having stored there on a sequence of instructions which when executed by a machine, causes the machine to perform a method comprising:

collecting data comprising a plurality of identification metadata and for each of the identification metadata, a corresponding voiceprint;

comparing each of the collected identification metadata to fraud data in a fraud report;

identifying which of the compared identification metadata are associated with the fraud data;

generating a blacklist comprising a list of voiceprints corresponding to identification metadata that are identified as associated with fraud data;

generating a whitelist comprising a list of voiceprints corresponding to identification metadata that are not identified as associated with fraud data; and

processing requests to screen speakers against one of the blacklist and the whitelist.

12. The computer-readable storage medium of claim 11, wherein the method further comprises indexing the collected data based on identification metadata that at least partially identifies a speaker associated with the data.

13. The computer-readable storage medium of claim 11, wherein the identifying comprises marking data associated with fraud if a fraudulent transaction can be linked to the voiceprint based on a fraud report.

14. The computer-readable storage medium of claim 13, wherein the generating a whitelist comprises saving each collected voiceprint not identified as associated with fraud to a whitelist database.

15. The computer-readable storage medium of claim 14, wherein the saving is performed only after effluxion of a predefined period of time after the collection of the collected voiceprint provided no fraudulent transaction was linked to the collected voiceprint during the predefined period.

16. The computer-readable storage medium of claim 14, wherein the saving is performed only after a number of non-fraudulent transactions linked to the collected voiceprint exceeds a threshold.

17. The computer-readable storage medium of claim 11, wherein the method further comprises generating a blacklist comprising data corresponding to the collected data identified as associated with fraud.