DEVICE, SYSTEM, AND METHOD OF LIVENESS DETECTION UTILIZING VOICE BIOMETRICS

• US 20150046162A1
• Filed: 10/27/2014
• Published: 02/12/2015
• Est. Priority Date: 11/26/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Krause, Nir Moshe, Salmon, Michael Itzhak, Gazit, Ran Yehoshua, ALEY-RAZ, ALMOG
• Granted Patent: US 9,484,037 B2
• Time in Patent Office: Days
• US Class Current: 704/247
• CPC Class Codes: G10L 17/04 Training, enrolment or model building
• CPC Class Codes: G10L 17/10 Multimodal systems, i.e. based on the integration of multiple recognition engines or fusion of expert systems
• CPC Class Codes: G10L 17/24 the user being prompted to utter a password or a predefined phrase

Abstract

Device, system, and method of liveness detection using voice biometrics. For example, a method comprises: generating a first matching score based on a comparison between: (a) a voice-print from a first text-dependent audio sample received at an enrollment stage, and (b) a second text-dependent audio sample received at an authentication stage; generating a second matching score based on a text-independent audio sample; and generating a liveness score by taking into account at least the first matching score and the second matching score.

Claims

1-20. -20. (canceled)

21. A method for performing an authentication evaluation for a user, the method comprising:

generating, with a processor, a first matching score based on a comparison between;

a voice-print from a first text-dependent audio sample received at an enrollment stage, and a second text-dependent audio sample received from the user at an evaluation stage;

subsequent to the comparison between the voice-print from the first text-dependent audio sample and the second text-dependent audio sample, generating, with a processor, a second matching score based on a comparison between data corresponding to a text-independent audio sample received from the user at the evaluation stage and data corresponding to the second text-dependent audio sample received from the user at the evaluation stage; and

making an authentication decision about the user based on at least the first matching score and the second matching score.

22. The method of claim 21, further comprising:

generating, with a processor, an utterance validation score based, at least in part, on the text-independent audio sample received at the evaluation stage; and

determining, based on the utterance validation score, whether the text-independent audio sample comprises an utterance of a phrase, wherein the user is prompted to utter the phrase at the evaluation stage.

23. The method of claim 22, wherein generating the utterance validation score comprises:

performing automatic speech recognition on the text-independent audio sample received at the evaluation stage.

24. The method of claim 21, wherein generating the second matching score based on a comparison between data corresponding to the text-independent audio sample and data corresponding to the second text-dependent audio sample comprises:

comparing a voice-print from the second text-dependent audio sample received at the evaluation stage, and the text-independent audio sample received at the evaluation stage.

25. The method of claim 21, wherein generating the second matching score based on a comparison between data corresponding to the text-independent audio sample and data corresponding to the second text-dependent audio sample comprises:

generating a first interim score based on a comparison between the text-independent audio sample received at the evaluation stage, and a text-independent voice-print of the second text-dependent audio sample received at the evaluation stage;

generating a second interim score based on a comparison between the second text-dependent audio sample received at the evaluation stage, and a text-independent voice-print of the text-independent audio sample received at the evaluation stage; and

generating the second matching score based, at least in part, on the first interim score and the second interim score.

26. The method of claim 21, wherein the first text-dependent audio received at the enrollment stage comprises speech of one or more phrases, and wherein the method further comprises:

prompting the user to repeat at least one of the one or more phrases in the evaluation stage; and

recording speech uttered by the user in response to prompting the user to repeat at least one of the one or more phrases, wherein the second text-dependent audio sample received at the evaluation stage comprises the recorded speech uttered in response to prompting the user to repeat at least one of the one or more phrases.

27. The method of claim 26, wherein the method further comprises:

prompting the user to utter a phrase, in the evaluation stage, different from the one or more phrases included in the first text-dependent audio received at the enrollment stage; and

recording speech uttered by the user in response to prompting the user to utter the phrase, wherein the second text-dependent audio sample received at the evaluation stage comprises the recorded speech uttered in response to prompting the user to utter the phrase.

28. The method of claim 21, wherein the enrollment stage is performed over a telephonic communication channel and/or the evaluation stage is performed over a telephonic communication channel.

29. An apparatus comprising:

at least one processor; and

at least one machine-readable medium configured to store instructions, which, when executed by the at least one processor, cause the apparatus to perform an authentication evaluation for a user, wherein performing the authentication evaluation comprises;

generating a first matching score based on a comparison between;

a voice-print from a first text-dependent audio sample received at an enrollment stage, and a second text-dependent audio sample received from the user at an evaluation stage;

subsequent to the comparison between the voice-print from the first text-dependent audio sample and the second text-dependent audio sample, generating a second matching score based on a comparison between data corresponding to a text-independent audio sample received from the user at the evaluation stage and data corresponding to the second text-dependent audio sample received from the user at the evaluation stage; and

making an authentication decision about the user based on at least the first matching score and the second matching score.

30. The apparatus of claim 29, wherein performing the authentication evaluation further comprises:

generating an utterance validation score based, at least in part, on the text-independent audio sample received at the evaluation stage,

wherein generating the utterance validation score comprises performing automatic speech recognition on the text-independent audio sample received at the evaluation stage.

31. The apparatus of claim 30, wherein generating the utterance validation score comprises:

performing automatic speech recognition on the text-independent audio sample received at the evaluation stage.

32. The apparatus of claim 29, wherein generating the second matching score based on a comparison between data corresponding to the text-independent audio sample and data corresponding to the second text-dependent audio sample comprises:

comparing a voice-print from the second text-dependent audio sample received at the evaluation stage, and the text-independent audio sample received at the evaluation stage.

33. The apparatus of claim 29, wherein generating the second matching score based on a comparison between data corresponding to the text-independent audio sample and data corresponding to the second text-dependent audio sample comprises:

generating a first interim score based on a comparison between the text-independent audio sample received at the evaluation stage, and a text-independent voice-print of the second text-dependent audio sample received at the evaluation stage;

generating a second interim score based on a comparison between the second text-dependent audio sample received at the evaluation stage, and a text-independent voice-print of the text-independent audio sample received at the evaluation stage; and

generating the second matching score based, at least in part, on the first interim score and the second interim score.

34. The apparatus of claim 29, wherein the first text-dependent audio received at the enrollment stage comprises speech of one or more phrases, and wherein performing the authentication evaluation further comprises:

prompting the user to repeat at least one of the one or more phrases in the evaluation stage; and

recording speech uttered by the user in response to prompting the user to repeat at least one of the one or more phrases, wherein the second text-dependent audio sample received at the evaluation stage comprises the recorded speech uttered in response to prompting the user to repeat at least one of the one or more phrases.

35. The apparatus of claim 34, wherein the performing the authentication evaluation further comprises:

prompting the user to utter a phrase, in the evaluation stage, different from the one or more phrases included in the first text-dependent audio received at the enrollment stage; and

recording speech uttered by the user in response to prompting the user to utter the phrase, wherein the second text-dependent audio sample received at the evaluation stage comprises the recorded speech uttered in response to prompting the user to utter the phrase.

36. At least one non-transitory machine-readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform a method comprising:

generating, with a processor, a first matching score based on a comparison between;

a voice-print from a first text-dependent audio sample received at an enrollment stage, and a second text-dependent audio sample received from the user at an evaluation stage;

subsequent to the comparison between the voice-print from the first text-dependent audio sample and the second text-dependent audio sample, generating, with a processor, a second matching score based on a comparison between data corresponding to a text-independent audio sample received from the user at the evaluation stage and data corresponding to the second text-dependent audio sample received from the user at the evaluation stage; and

making an authentication decision about the user based on at least the first matching score and the second matching score.

37. The at least one non-transitory machine-readable medium of claim 36, wherein the method further comprises:

generating, with a processor, an utterance validation score based, at least in part, on the text-independent audio sample received at the evaluation stage; and

determining, based on the utterance validation score, whether the text-independent audio sample comprises an utterance of a phrase, wherein the user is prompted to utter the phrase at the evaluation stage.

38. The at least one non-transitory machine-readable medium of claim 37, wherein generating the utterance validation score comprises:

performing automatic speech recognition on the text-independent audio sample received at the evaluation stage.

39. The at least one non-transitory machine-readable medium of claim 36, wherein generating the second matching score based on a comparison between data corresponding to the text-independent audio sample and data corresponding to the second text-dependent audio sample comprises:

comparing a voice-print from the second text-dependent audio sample received at the evaluation stage, and the text-independent audio sample received at the evaluation stage.

40. The at least one non-transitory machine-readable medium of claim 36, wherein generating the second matching score based on a comparison between data corresponding to the text-independent audio sample and data corresponding to the second text-dependent audio sample comprises:

generating a first interim score based on a comparison between the text-independent audio sample received at the evaluation stage, and a text-independent voice-print of the second text-dependent audio sample received at the evaluation stage;

generating a second interim score based on a comparison between the second text-dependent audio sample received at the evaluation stage, and a text-independent voice-print of the text-independent audio sample received at the evaluation stage; and

generating the second matching score based, at least in part, on the first interim score and the second interim score.