Directory dialer name recognition

• US 8,369,492 B2
• Filed: 07/07/2008
• Issued: 02/05/2013
• Est. Priority Date: 11/26/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Janke, Eric William, Sloan, Keith
• Primary Examiner(s): Gauthier, Gerald
• Application Number:
  US12/168,444
• Publication Number:
  US 20080267364A1
• Time in Patent Office: 1,674 Days
• Field of Search: 379/88.01, 379/88.03, 379/88.04, 379/88.16, 379/433.07, 379/93.15, 379/399.01, 704/244, 704/254, 704/277, 704/231, 704/240, 704/243, 704/251, 704/256.1, 704/270, 704/275, 370/352, 455/564, 710/1
• US Class Current: 379/88.03
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: H04M 3/42204 Arrangements at the exchange for service or number selection by voice at the terminal H04M1/27

Abstract

A method, apparatus, computer program product and service for directory dialer name recognition. The directory dialer has a directory of names and a first name grammar and a second name grammar representing phonetic baseforms of first names and second names respectively. The method includes: receiving voice data for a spoken name after requesting a user to speak the required name; extracting a set of phonetic baseforms for the voice data; and finding the best matches between the extracted set of phonetic baseforms voice data and any combination of the first name grammar and the second name grammar. The method can further include: checking the best match against the directory of names; if the best match does not exist in the directory, informing the user and prompting the next best match as an alternative; and if the best match does exist in the directory, forwarding the call to that best match.

Claims

1. An interactive voice response system comprising:

a database storing a directory of names, each name specifying a first name and a second name; and

a server, capable of accessing the database, the server configured to perform;

receiving voice data for a spoken name from a user;

extracting a set of phonetic baseforms from the voice data;

finding a best match to the extracted set of phonetic baseforms from combinations of phonetic baseforms from a first name grammar representing phonetic baseforms for first names in the directory of names and a second name grammar representing phonetic baseforms for second names in the directory of names;

determining whether a combination of first name and second name found to be the best match exists in the directory of names; and

forwarding the call if the combination is determined to exist in the directory of names.

2. The interactive voice response system according to claim 1, the the server further configured to perform:

if the combination of first name and last name found to be the best match is determined not to exist in the directory of names, informing the user and prompting the user with an alternative combination of first name and second name found to be a next best match.

3. The interactive voice response system according to claim 1, wherein finding the best match comprises:

finding a first set of best first name matches to the extracted set of phonetic baseforms from a set of phonetic baseforms in the first name grammar;

finding a second set of best second name matches to the extracted set of phonetic baseforms from a set of phonetic baseforms in the second name grammar;

creating a list of combinations of phonetic baseforms in the first set and the second set; and

selecting the combination with a highest confidence score as the best match.

4. The interactive voice response system according to claim 1, wherein the first name grammar is derived from phonemes for first names in the directory and the second name grammar is derived from phonemes for second names in the directory.

5. The interactive voice response system according to claim 1, wherein the directory indicates which names are included in the first and second name grammars so that the first and second name grammars are updated only with new first names and new second names in the directory.

6. The interactive voice response system according to claim 1, wherein first and second name grammar memories comprise first and second names not included in the directory.

7. A computer-readable storage medium storing computer program instructions that, when executed by a computer, cause the computer to perform:

receiving voice data for a spoken name from a user;

extracting a set of phonetic baseforms from the voice data;

finding a best match to the extracted set of phonetic baseforms from combinations of phonetic baseforms from a first name grammar representing phonetic baseforms for first names in a directory of names and a second name grammar representing phonetic baseforms for second names in the directory of names;

determining whether a combination of first name and second name found to be the best match exists in the directory of names; and

forwarding the call if the combination is determined to exist in the directory of names.

8. The computer readable storage medium according to claim 7, wherein the computer program instructions further cause the computer to perform:

if the combination of first name and last name found to be the best match is determined not to exist in the directory of names, informing the user and prompting the user with an alternative combination of first name and second name found to be a next best match.

9. The computer readable storage medium according to claim 7, wherein finding the best match comprises:

finding a first set of best first name matches to the extracted set of phonetic baseforms from a set of phonetic baseforms in the first name grammar;

finding a second set of best second name matches to the extracted set of phonetic baseforms from a set of phonetic baseforms in the second name grammar;

creating a list of combinations of phonetic baseforms in the first set and the second set; and

selecting the combination with a highest confidence score as the best match.

10. The computer readable storage medium according to claim 7, wherein the first name grammar is derived from phonemes for first names in the directory and the second name grammar is derived from phonemes for second names in the directory.

11. The computer readable storage medium according to claim 7, wherein the directory indicates which names are included in the first and second name grammars so that the first and second name grammars are updated only with new first names and new second names in the directory.

12. The computer readable storage medium according to claim 7, wherein first and second name grammar memories comprise first and second names not included in the directory.