System and method for detecting synthetic speaker verification

• US 9,812,133 B2
• Filed: 08/05/2016
• Issued: 11/07/2017
• Est. Priority Date: 04/11/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Schroeter, Horst J.
• Primary Examiner(s): COLUCCI, MICHAEL C
• Application Number:
  US15/229,376
• Publication Number:
  US 20160343379A1
• Time in Patent Office: 459 Days
• Field of Search: 704246, 704275, 704273, 704271, 704252, 704250, 704249, 704248, 704247, 704219, 704208, 463 29, 455411, 455406, 455563, 379361, 713186, 713169
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 17/04 Training, enrolment or model building
• CPC Class Codes: G10L 17/08 Use of distortion metrics or a particular distance between probe pattern and reference templates
• CPC Class Codes: G10L 17/20 Pattern transformations or operations aimed at increasing system robustness, e.g. against channel noise or different working conditions
• CPC Class Codes: G10L 17/24 the user being prompted to utter a password or a predefined phrase

Abstract

Disclosed herein are systems, methods, and tangible computer readable-media for detecting synthetic speaker verification. The method comprises receiving a plurality of speech samples of the same word or phrase for verification, comparing each of the plurality of speech samples to each other, denying verification if the plurality of speech samples demonstrate little variance over time or are the same, and verifying the plurality of speech samples if the plurality of speech samples demonstrates sufficient variance over time. One embodiment further adds that each of the plurality of speech samples is collected at different times or in different contexts. In other embodiments, variance is based on a pre-determined threshold or the threshold for variance is adjusted based on a need for authentication certainty. In another embodiment, if the initial comparison is inconclusive, additional speech samples are received.

Claims

1. A method comprising:

receiving, via a processor and from a user, a plurality of speech samples of a same word, wherein the plurality of speech samples of the same word comprises a current speech sample and a plurality of previously recorded speech samples;

generating, via the processor, a sample similarity from the plurality of speech samples;

making a decision, via the processor, whether to enroll the user in a speech verification system according to a comparison of the sample similarity with a threshold, wherein the decision is a first decision if the sample similarity is above the threshold and the decision is a second decision different from and mutually exclusive with the first decision if the sample similarity is below the threshold; and

enrolling the user in the speech verification system based at least in part on the decision.

2. The method of claim 1, wherein the first sample of the plurality of speech samples and the second sample in the plurality of speech samples are each associated with a distinct location.

3. The method of claim 1, wherein the comparison comprises a comparison of each of the first sample in the plurality of speech samples and the second sample in the plurality of speech samples with other samples of the plurality of speech samples.

4. The method of claim 1, wherein the comparison results in a variance that identifies how different the first sample is from the second sample.

5. The method of claim 4, wherein when the variance is below a threshold, then not enrolling the user in the speech verification system.

6. The method of claim 5, wherein when the variance is at or above the threshold, then enrolling the user in the speech verification system.

7. The method of claim 4, wherein the variance has a range which varies based on a job of the user.

8. A system comprising:

a processor; and

a computer-readable storage medium having instructions stored which, when executed by the processor, result in the processor performing operations comprising;

receiving, via a processor and from a user, a plurality of speech samples of a same word, wherein the plurality of speech samples of the same word comprises a current speech sample and a plurality of previously recorded speech samples;

generating, via the processor, a sample similarity from the plurality of speech samples;

making a decision, via the processor, whether to enroll the user in a speech verification system according to a comparison of the sample similarity with a threshold, wherein the decision is a first decision if the sample similarity is above the threshold and the decision is a second decision different from and mutually exclusive with the first decision if the sample similarity is below the threshold; and

enrolling the user in the speech verification system based at least in part on the decision.

9. The system of claim 8, wherein the first sample of the plurality of speech samples and the second sample in the plurality of speech samples are each associated with a distinct location.

10. The system of claim 8, wherein the comparison comprises a comparison of each of the first sample in the plurality of speech samples and the second sample in the plurality of speech samples with other samples of the plurality of speech samples.

11. The system of claim 8, wherein the comparison results in a variance that identifies how different the first sample is from the second sample.

12. The system of claim 11, wherein when the variance is below a threshold, then not enrolling the user in the speech verification system.

13. The system of claim 12, wherein when the variance is at or above the threshold, then enrolling the user in the speech verification system.

14. The system of claim 11, wherein the variance has a range which varies based on a job of the user.

15. A computer-readable storage device having instructions stored which, when executed by a computing device, result in the computing device performing operations comprising:

receiving, via a processor and from a user, a plurality of speech samples of a same word, wherein the plurality of speech samples of the same word comprises a current speech sample and a plurality of previously recorded speech samples;

generating, via the processor, a sample similarity from the plurality of speech samples;

making a decision, via the processor, whether to enroll the user in a speech verification system according to a comparison of the sample similarity with a threshold, wherein the decision is a first decision if the sample similarity is above the threshold and the decision is a second decision different from and mutually exclusive with the first decision if the sample similarity is below the threshold; and

enrolling the user in the speech verification system based at least in part on the decision.

16. The computer-readable storage device of claim 15, wherein the first sample of the plurality of speech samples and the second sample in the plurality of speech samples are each associated with a distinct location.

17. The computer-readable storage device of claim 15, wherein the comparison comprises a comparison of each of the first sample in the plurality of speech samples and the second sample in the plurality of speech samples with other samples of the plurality of speech samples.

18. The computer-readable storage device of claim 15, wherein the comparison results in a variance that identifies how different the first sample is from the second sample.

19. The computer-readable storage device of claim 18, wherein when the variance is below a threshold, then not enrolling the user in the speech verification system.

20. The computer-readable storage device of claim 19, wherein when the variance is at or above the threshold, then enrolling the user in the speech verification system.