Speaker verification method

• US 20050143996A1
• Filed: 01/12/2005
• Published: 06/30/2005
• Est. Priority Date: 01/21/2000
• Status: Active Grant

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Bossemeyer, Robert Wesley Jr., Ratasuk, Rapeepat
• Granted Patent: US 7,630,895 B2
• Time in Patent Office: Days
• US Class Current: 704/247
• CPC Class Codes: G10L 17/06 Decision making techniques; Pattern matching strategies
• CPC Class Codes: G10L 17/08 Use of distortion metrics or a particular distance between probe pattern and reference templates

Abstract

A speaker verification method consist of the following steps: (1) generating a code book (42) covering a number of speakers having a number of training utterances for each of the speakers; (2) receiving a number of test utterances (44) from a speaker; (3) comparing (46) each of the test utterances to each of the training utterances for the speaker to form a number of decisions, one decision for each of the number of test utterances; (4) weighting each of the decisions (48) to form a number of weighted decisions; and (5) combining (50) the plurality of weighted decisions to form a verification decision (52).

Claims

1-21. -21. (canceled)

22. A speaker verification method comprising the steps of:

(a) comparing each of a Plurality of test utterances to a plurality of training utterances for a speaker to form a Plurality of preliminary verification decisions, one preliminary verification decision of the plurality of preliminary verification decisions for each of the plurality of test utterances, wherein the one preliminary verification decision is either a true or a false decision; and

(b) combining the plurality of preliminary verification decisions to form a verification decision.

23. The method of claim 22, wherein step (a) further includes the steps of:

(a1) generating a code book covering a plurality of speakers having the plurality of training utterances for each of the plurality of speakers, wherein a first utterance of the Plurality of training utterances is not the same as a second utterance of the plurality of training utterances.

24. The method of claim 23, further including the step of:

(a2) receiving the plurality of test utterances from one of the plurality of speakers, wherein a first of the plurality of test utterances is not the same as a second of the plurality of test utterances;

25. The method of claim 22, wherein step (b) further includes the steps of:

(b1) weighting each of the Plurality of Preliminary verification decisions.

26. The method of claim 25, wherein the step of weighting includes determining a historical probability of false alarm for each of a Plurality of utterances.

27. The method of claim 26, wherein the step of weighting includes evaluating a quality of the Preliminary verification decision for each of the plurality of preliminary verification decisions.

29. A method of sneaker verification, comprising the steps:

(a) when the sneaker is male, using a male variance vector to determine a weighted Euclidean distance between each of a plurality of test utterances and the plurality of training utterances;

(b) forming a preliminary verification decision for each of the plurality of test utterances to form a plurality of preliminary verification decisions;

(c) combining the plurality of preliminary verification decisions to form a verification decision.

30. The method of claim 29, wherein step (a) further includes the steps of:

(a1) receiving the plurality of test utterances from a speaker.

31. The method of claim 30, further including the step of:

(a2) determining if the speaker is the male;

32. The method of claim 29, further including the steps of:

(d) when the speaker is not male, using a female variance vector to determine a weighted Euclidean distance between each of the Plurality of test utterances and the plurality of training utterances;

(e) forming a Preliminary verification decision for each of the plurality of test utterances to form a Plurality of Preliminary verification decisions;

(f) combining the Plurality of preliminary verification decisions to from a verification decision.

33. A speaker verification method comprising the steps of:

(a) comparing a Plurality of different test utterances to a Plurality of training utterances for the speaker to form a Plurality of Preliminary verification decisions, one Preliminary verification decision for each of the plurality of different test utterances;

(b) combining the Plurality of Preliminary verification decisions to form a verification decision.

34. The method of claim 33, wherein step (a) further includes the step of:

(a1) receiving a plurality of different test utterances from a speaker;