Method and apparatus for generating and updating a voice tag

• US 7,471,775 B2
• Filed: 06/30/2005
• Issued: 12/30/2008
• Est. Priority Date: 06/30/2005
• Status: Expired due to Fees

• Current Assignee: Google Technology Holdings LLC (Alphabet Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Cheng, Yan Ming
• Primary Examiner(s): Gauthier; Gerald
• Application Number:
  US11/170,892
• Publication Number:
  US 20070019793A1
• Time in Patent Office: 1,279 Days
• Field of Search: 379/88.01, 704/240, 704/249, 704/251, 704/270.1, 704/247, 455/563
• US Class Current: 379/88.01
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting
• CPC Class Codes: H04M 1/271 controlled by voice recognition
• CPC Class Codes: H04M 2250/74 with voice recognition means voice activated dialling H04M1/271; voice recognition algorithms G10L15/00

Abstract

A method and apparatus (100) for updating a voice tag comprising N stored voice tag phoneme sequences includes a function (110) for determining (205) an accepted stored voice tag phoneme sequence for an utterance, a function (140) for extracting(210) a current set of M phoneme sequences having highest likelihoods of representing the utterance, a function (160) for updating (215) a reference histogram associated with the accepted voice tag, and a function (160) for updating (225) the voice tag with N selected phoneme sequences that are selected from the current set of M phoneme sequences and the set of N voice tag phoneme sequences, wherein the N selected phoneme sequences have phoneme histograms most closely matching the reference histogram. The method and apparatus (100) also generates a voice tag using some functions (110, 140, 160) that are common with the method and apparatus to update the voice tag, such as the extracting (410) of the current set of M phoneme sequences.

Claims

1. A method used to update a voice tag, the method comprising:

determining an accepted stored voice tag from an utterance, wherein the accepted stored voice tag comprises a stored set of N voice tag phoneme sequences that are associated with a semantic and a reference phoneme histogram;

extracting a current set of M phoneme sequences having highest likelihoods of representing the utterance, using a stored set of language phonemes;

updating the reference phoneme histogram by using an updating phoneme histogram derived from the current set of M phoneme sequences; and

updating the voice tag with N selected phoneme sequences that are selected from the current set of M phoneme sequences and the stored set of N voice tag phoneme sequences, wherein the N selected phoneme sequences have associated phoneme histograms most closely matching the reference phoneme histogram, and wherein the updating of the voice tag is performed after the updating of the reference phoneme histogram.

2. The method according to claim 1, wherein the updating of the reference phoneme histogram further comprises generating the updating phoneme histogram as an n-gram histogram from a phoneme sequence of the current set of phoneme sequences that has a highest likelihood of representing the utterance.

3. The method according to claim 1 further comprising:

comparing an n-gram phoneme histogram of each phoneme sequence of the current set of M phoneme sequences and of each phoneme sequence of the stored set of N voice tag phoneme sequences to the reference phoneme histogram to determine M+N histogram matching metrics.

4. The method according to claim 1, wherein the stored set of language phonemes is for a particular language.

5. The method according to claim 1, wherein the stored set of language phonemes is a set of speaker independent phonemes.

6. The method according to claim 1, further comprising storing the voice tag in association with a semantic value.

7. The method according to claim 1, wherein the extracting of the current set of M phoneme sequences comprises using a hidden Markov model.

8. A method used to generate a voice tag, the method comprising:

determining a command to generate a voice tag from an utterance;

extracting a current set of phoneme sequences having highest likelihoods of representing the utterance, using a set of stored language phonemes;

generating a reference phoneme histogram that comprises an n-gram phoneme histogram of the phoneme sequence of the current set of phoneme sequences that has a highest likelihood of representing the utterance; and

storing up to N phoneme sequences selected from the current set of phoneme sequences as a voice tag, in association with the reference phoneme histogram.

9. The method according to claim 8, further comprising:

determining no existence of an acceptable stored voice tag phoneme sequence for the utterance before determining the command to generate a voice tag from the utterance.

10. The method according to claim 8 further comprising:

storing semantic information in association with the voice tag.

11. An electronic device comprising:

means for determining an accepted voice tag phoneme sequence from an utterance, wherein the accepted voice tag phoneme sequence is in a stored set of N voice tag phoneme sequences that form a voice tag that is associated with a semantic and a reference phoneme histogram;

means for extracting a current set of M phoneme sequences having highest likelihoods of representing the utterance, using a stored set of language phonemes;

means for updating the reference phoneme histogram that uses an updating phoneme histogram derived from the current set of M phoneme sequences; and

means for updating the voice tag with N selected phoneme sequences that are selected from the current set of M phoneme sequences and the stored set of N voice tag phoneme sequences, wherein the N selected phoneme sequences have associated phoneme histograms most closely matching the reference phoneme histogram.