Speaker verification methods and apparatus

• US 8,620,657 B2
• Filed: 09/14/2012
• Issued: 12/31/2013
• Est. Priority Date: 10/24/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Farrell, Kevin R., James, David A., Ganong, William F. III, Carter, Jerry K.
• Primary Examiner(s): Chawan, Vijay B
• Application Number:
  US13/617,196
• Publication Number:
  US 20130030809A1
• Time in Patent Office: 473 Days
• Field of Search: 704/270, 704/273, 704/246, 704/275, 704/254, 704503-504, 704/231, 379/88.03, 379/88.18, 379/93.06, 379/88.02, 379/88.01, 379/189, 455/557, 455/411, 381/94.9, 381/386, 340/7.57, 340/990, 340/995.12
• US Class Current: 704/246
• CPC Class Codes: G07C 9/37 using biometric data, e.g. fingerprints, iris scans or voice recognition
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: G10L 17/24 the user being prompted to utter a password or a predefined phrase

Abstract

One aspect includes determining validity of an identity asserted by a speaker using a voice print associated with a user whose identity the speaker is asserting, the voice print obtained from characteristic features of at least one first voice signal obtained from the user uttering at least one enrollment utterance including at least one enrollment word by obtaining a second voice signal of the speaker uttering at least one challenge utterance that includes at least one word not in the at least one enrollment utterance, obtaining at least one characteristic feature from the second voice signal, comparing the at least one characteristic feature with at least a portion of the voice print to determine a similarity between the at least one characteristic feature and the at least a portion of the voice print, and determining whether the speaker is the user based, at least in part, on the similarity.

Claims

1. A method for determining validity of an identity asserted by a speaker using a voice print associated with a user whose identity the speaker is asserting, the voice print obtained from the user'"'"'s utterance of at least one enrollment utterance including at least one enrollment word, the method comprising:

obtaining a voice signal of the speaker uttering at least one challenge utterance, wherein the at least one challenge utterance includes at least one word that was not in the at least one enrollment utterance; and

determining whether the speaker is the user based, at least in part, on the voice signal and the voice print.

2. The method of claim 1, wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance consisting preponderantly of words not used in the at least one enrollment utterance.

3. The method of claim 2, wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance consisting substantially of words not used in the at least one enrollment utterance.

4. The method of claim 1, wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance selected from a challenge vocabulary comprising a plurality of challenge words, wherein the challenge vocabulary includes words having substantial phonetic overlap with the at least one word used in the at least one enrollment utterance.

5. The method of claim 4, wherein each of the plurality of words in the challenge vocabulary has at least one syllable that rhymes with at least one syllable of at least one word used in the at least one enrollment utterance.

6. The method of claim 1, wherein the at least one enrollment word is selected from an enrollment vocabulary, and wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance selected from a challenge vocabulary comprising a plurality of challenge words, and wherein the challenge vocabulary includes more words than the enrollment vocabulary.

7. The method of claim 6, wherein each distinct word used in the at least one enrollment utterance has a plurality of corresponding words in the challenge vocabulary, each having phonetic overlap with the corresponding word in the at least one enrollment utterance.

8. The method of claim 1, wherein the challenge vocabulary includes at least 25 words from which the at least one challenge utterance may be formed.

9. The method of claim 1, wherein the challenge vocabulary includes at least 50 words from which the at least one challenge utterance may be formed.

10. The method of claim 1, further comprising performing automatic speech recognition on the voice signal of the speaker to verify that the speaker spoke the at least one word in the at least one challenge utterance.

11. At least one computer readable medium encoded with at least one program for execution on at least one processor, the program having instructions that, when executed by the at least one processor, perform a method of determining a validity of an identity asserted by a speaker using a voice print associated with a user whose identity the speaker is asserting, the voice print obtained from the user'"'"'s utterance of at least one enrollment utterance including at least one enrollment word, the method comprising:

obtaining a voice signal of the speaker uttering at least one challenge utterance, wherein the at least one challenge utterance includes at least one word that was not in the at least one enrollment utterance; and

determining whether the speaker is the user based, at least in part, on the voice signal and the voice print.

12. The at least one computer readable medium of claim 11, wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance consisting preponderantly of words not used in the at least one enrollment utterance.

13. The at least one computer readable medium of claim 12, wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance consisting substantially of words not used in the at least one enrollment utterance.

14. The at least one computer readable medium of claim 11, wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance selected from a challenge vocabulary comprising a plurality of challenge words, wherein the challenge vocabulary includes words having substantial phonetic overlap with the at least one enrollment word used in the at least one enrollment utterance.

15. The at least one computer readable medium of claim 14, wherein each of the plurality of words in the challenge vocabulary has at least one syllable that rhymes with at least one syllable of at least one word used in the at least one enrollment utterance.

16. The at least one computer readable medium of claim 11, wherein the at least one enrollment word is selected from an enrollment vocabulary, and wherein obtaining the voice signal of the speaker uttering the at least one challenge utterance includes obtaining at least one challenge utterance selected from a challenge vocabulary comprising a plurality of challenge words, and wherein the challenge vocabulary includes more words than the enrollment vocabulary.

17. The at least one computer readable medium of claim 16, wherein each distinct word used in the at least one enrollment utterance has a plurality of corresponding words in the challenge vocabulary, each having phonetic overlap with the corresponding word in the at least one enrollment utterance.

18. The at least one computer readable medium of claim 11, wherein the challenge vocabulary includes at least 25 words from which the at least one challenge utterance may be formed.

19. The at least one computer readable medium of claim 11, wherein the challenge vocabulary includes at least 50 words from which the at least one challenge utterance may be formed.

20. The at least one computer readable medium of claim 11, the method further comprising performing automatic speech recognition on the voice signal of the speaker to verify that the speaker spoke the at least one word in the at least one challenge utterance.

21. A speaker verification system comprising:

at least one computer readable storage medium storing a voice print obtained from a user'"'"'s utterance of at least one enrollment utterance;

a receiver to receive a voice signal of the speaker uttering at least one challenge utterance, wherein the at least one challenge utterance includes at least one word that was not in the at least one enrollment utterance; and

at least one controller, coupled to the memory and the receiver, configured to determine whether the speaker is the user based, at least in part, on voice signal and the voice print.

22. The speaker verification system of claim 21, wherein the receiver receives the voice signal of the speaker uttering the at least one challenge utterance consisting preponderantly of words not used in the at least one enrollment utterance.

23. The speaker verification system of claim 22, wherein the receiver receives the voice signal of the speaker uttering the at least one challenge utterance consisting substantially of words not used in the at least one enrollment utterance.

24. The speaker verification system of claim 21, wherein the receiver receives the voice signal of the speaker uttering at least one challenge utterance selected from a challenge vocabulary comprising a plurality of challenge words, wherein the challenge vocabulary includes words having substantial phonetic overlap with the at least one enrollment word used in the at least one enrollment utterance.

25. The speaker verification system of claim 24, wherein each of the plurality of words in the challenge vocabulary has at least one syllable that rhymes with at least one syllable of at least one word used in the at least one enrollment utterance.

26. The speaker verification system of claim 21, wherein the at least one enrollment word is selected from an enrollment vocabulary, and wherein the receiver receives the voice signal of the speaker uttering at least one challenge utterance selected from a challenge vocabulary comprising a plurality of challenge words, and wherein the challenge vocabulary includes more words than the enrollment vocabulary.

27. The speaker verification system of claim 26, wherein each distinct word used in the at least one enrollment utterance has a plurality of corresponding words in the challenge vocabulary, each having phonetic overlap with the corresponding word in the at least one enrollment utterance.

28. The speaker verification system of claim 21, wherein the challenge vocabulary includes at least 25 words from which the at least one challenge utterance may be formed.

29. The speaker verification system of claim 26, wherein the challenge vocabulary includes at least 50 words from which the at least one challenge utterance may be formed.

30. The speaker verification system of claim 26, wherein the controller is configured to perform automatic speech recognition on the voice signal of the speaker to verify that the speaker spoke the at least one word in the at least one challenge utterance.