METHOD FOR VERIFYING THE IDENTITYOF A SPEAKER, SYSTEM THEREFORE AND COMPUTER READABLE MEDIUM

• US 20120173239A1
• Filed: 06/26/2009
• Published: 07/05/2012
• Est. Priority Date: 12/10/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Agnitio s.l. (Microsoft Corporation)
• Inventors: Garcia Gomar, Marta, Sánchez Asenjo, Marta
• Granted Patent: US 9,792,912 B2
• Time in Patent Office: Days
• US Class Current: 704/246
• CPC Class Codes: B66B 13/26 between closing doors
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 17/02 Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction
• CPC Class Codes: G10L 17/06 Decision making techniques; Pattern matching strategies
• CPC Class Codes: G10L 17/24 the user being prompted to utter a password or a predefined phrase
• CPC Class Codes: G10L 17/26 Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices

Abstract

The invention refers to a method of verifying the identity of a speaker based on the speakers voice comprising the steps of: receiving (1, 5) a first and a second voice utterance; using biometric voice data to verify (2, 6) that the speakers voice corresponds to the speaker the identity of which is to be verified based on the received first and/or second voice utterance and determine (8) the similarity of the two received voice utterances characterized in that the similarity is determined using biometric voice characteristics of the two voice utterances or data derived from such biometric voice characteristics. The invention further refers to a System (80) for verifying the identity of a speaker based on the speakers voice comprising: a component (81) for receiving a first and a second voice utterance; a component (82) for using biometric voice data to verify that the speakers voice corresponds to the speaker the identity of which is to be verified based on the received first and/or second voice utterance and a component (83) for comparing the two received voice utterances in order to determine the similarity of the two voice utterances characterized in that the similarity is determined using biometric voice characteristics of the two voice utterances or data derived from such biometric voice characteristics.

Claims

1-14. -14. (canceled)

15. A method of verifying the identity of a speaker based on the speakers voice comprising the steps of:

a) receiving a first and a second voice utterance;

b) using biometric voice data to verify that the speakers voice corresponds to the speaker the identity of which is to be verified based on the received first and/or second voice utterance; and

c) determining the similarity of the two received voice utterances characterized in that the similarity is determined using biometric voice characteristics of the two voice utterances or data derived from such biometric voice characteristics.

16. The method of claim 15, wherein the second voice utterance is requested from the speaker after having used biometric voice data to verify that the speaker'"'"'s voice corresponds to the speaker the identity of which is to be verified;

and

wherein the speaker is requested to repeat the first voice utterance in order to receive the second voice utterance.

17. The method of claim 16, wherein biometric voice data are extracted from the two voice utterances and the same extracted data are used as the biometric voice data for verifying that the speaker'"'"'s voice corresponds to the speaker the identity of which is to be verified and as the biometric voice characteristics used for determining the similarity of the two voice utterances or for deriving data for determining the similarity of the two voice utterances.

18. The method of claim 15, wherein biometric voice characteristics or data derived therefrom are determined for different times such as, e.g. more than 100, 500, 1000 or 10,000 times within the two voice utterances for determining the similarity of the two voice utterances.

19. The method of claim 18, wherein the biometric voice characteristics used for determining the similarity of the two voice utterances comprise or are based on a set of at least n values such as a vector wherein each set of values is determined from a time slice of the voice utterance with a length of e.g. between 10 and 40 ms, wherein n is a number between 2 and 40.

20. The method of claim 19, wherein a set of values has at least 2 times n values wherein each set of values has n values as defined in claim 19 and n values resulting from the difference between the n values of the set of values and the corresponding n values of the temporarily previous or following set of values.

21. The method of claim 19, wherein a derived set of/values is calculated, preferably expressing a coincidence of a set of values of claim 19 with respect to a statistical voice model such as Gaussian Mixed Model, wherein the derived set of l values preferably expresses a coincidence of a set of values of claim 19 with respect to l statistical voice model components that define the statistical voice model.

22. The method of claim 21, wherein the statistical voice model is a statistical voice model particular for the speaker the identity of which is to be verified or a generic model independent of the speaker.

23. The method of claim 21, wherein multiple sets of values and/or multiple derived sets of values are calculated each of which corresponds to different time slices within a voice utterance.

24. The method of claim 23, wherein the multiple sets of values or multiple derived sets of values of two voice characteristics are compared using dynamic time warping, thereby preferably taking into account only a subset of each set of values of the multiple set of values or of the derived set of values of the multiple derived sets of values.

25. The method of claim 24, wherein the multiple sets of values or multiple derived sets of values of the two voice characteristics are compared using correlations calculated for a specific value or a specific subgroup of values of the sets of values or the derived sets of values.

26. The method of claim 15, wherein the identity of a speaker based on the speakers voice is indicated to be verified in case that:

using the biometric voice data to verify that the speakers voice corresponds to the speaker the identity of which is to be verified based on the received first and second voice utterance gives a positive result representing that the voice corresponds to the speaker the identity of which is to be verified; and

the two voice utterances are found to be not similar, e.g. by a measure of the similarity being below a threshold or the dissimilarity being above a threshold.

27. A computer-readable medium having instructions stored thereon, which when executed on a computer performs the method of claim 15.

28. A system for verifying the identity of a speaker based on the speakers voice, the system comprising:

a component for receiving a first and a second voice utterance;

a component for using biometric voice data to verify that the speakers voice corresponds to the speaker the identity of which is to be verified based on the received first and/or second voice utterance; and

a component for comparing the two received voice utterances in order to determine the similarity of the two voice utterances;

characterized in that the similarity is determined using biometric voice characteristics of the two voice utterances or data derived from such biometric voice characteristics.