Adapting a speech system to user pronunciation

• US 9,997,155 B2
• Filed: 09/09/2015
• Issued: 06/12/2018
• Est. Priority Date: 09/09/2015
• Status: Active Grant

• Current Assignee: GM Global Technology Operations LLC (General Motors Company)
• Original Assignee: GM Global Technology Operations LLC (General Motors Company)
• Inventors: Grost, Timothy J., Hansen, Cody R., Winter, Ute
• Primary Examiner(s): Sharma, Neeraj
• Application Number:
  US14/848,729
• Publication Number:
  US 20170069311A1
• Time in Patent Office: 1,007 Days
• Field of Search: 704201, 704 10, 704235, 704260, 704243
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting

Abstract

A system and method of adapting a speech system includes the steps of: receiving confirmation of a phonetic transcription of one or more names, receiving confirmation of a selected stored text result, and storing the phonetic transcription with the selected stored text result using an automatic speech recognition (ASR) system, a text-to-speech (TTS) system, or both.

Claims

1. A method of adapting speech systems, comprising the steps of:

a) receiving confirmation of a phonetic transcription of one or more names, wherein the phonetic transcription is a phonetic representation of a user'"'"'s pronunciation of the one or more names received as speech from the user;

b) generating, using an automatic speech recognition (ASR) system installed in the vehicle, a text interpretation of the confirmed phonetic transcription of the user'"'"'s pronunciation of the one or more names, wherein the ASR system includes a processor configured to execute instructions stored in a non-transitory memory of the ASR system;

c) using the ASR system, a text-to-speech (TTS) system at the vehicle, or both, selecting a stored text result by comparing the confirmed phonetic transcription associated with the text interpretation generated in step b), with phonetic transcriptions associated with stored text located within electronics on-board the vehicle or located within electronics at a remote location off-board the vehicle;

d) receiving confirmation of a selected stored text result based on the comparison in step c); and

e) storing the confirmed phonetic transcription with the selected stored text result in one or more modules, including memory modules, of the ASR system, the text-to-speech (TTS) system, or both, wherein the confirmed phonetic transcription represents a user-specific pronunciation of the selected stored text.

2. The method of claim 1, wherein the phonetic transcription is confirmed by the user, the ASR system, or the TTS system.

3. The method of claim 1, wherein the one or more names further comprises a name of a person, a song, a website, a file, a phone number, or a street address.

4. The method of claim 1, further comprising the step of initiating steps (a)-(e) when an output from the ASR system, the TTS system, or both falls below a predetermined confidence threshold.

5. The method of claim 1, wherein the ASR system is used with a remote speech recognition system accessed via a handheld wireless device.

6. A method of adapting speech systems, comprising the steps of:

a) receiving a spoken name at an automatic speech recognition (ASR) system installed in a vehicle from a user via a vehicle microphone, wherein the ASR system includes a processor configured to execute instructions stored in a non-transitory memory of the ASR system;

b) converting the spoken name into a phonetic transcription using the ASR system, wherein the phonetic transcription is a phonetic representation of the user'"'"'s pronunciation of the spoken name;

c) presenting the phonetic transcription to the user;

d) receiving confirmation at the vehicle that the phonetic transcription is accurate;

e) selecting text representing the spoken name;

f) generating a phonetic interpretation from the selected text in step (e), wherein the phonetic interpretation is a phonetic representation of the selected text;

g) comparing the phonetic interpretation of the selected text in step f) with the confirmed phonetic transcription of the user'"'"'s pronunciation of the spoken name in step (d); and

h) storing the confirmed phonetic transcription in one or more modules, including memory modules, of the ASR system, a text-to-speech (TTS) system, or both, based on the comparison in step g), wherein the confirmed phonetic transcription represents a user-specific pronunciation of the selected text.

7. The method of claim 6, wherein a text interpretation is generated from the confirmed phonetic transcription in step (d).

8. The method of claim 6, wherein the phonetic transcription is confirmed by the user, the ASR system, or the TTS system.

9. The method of claim 6, wherein the spoken name further comprises a name of a person, a song, a website, a file, a phone number, or a street address.

10. The method of claim 6, further comprising the step of initiating steps (a)-(h) when the output from the ASR system, the TTS system, or both falls below a predetermined confidence threshold.

11. The method of claim 6, wherein the text in step (e) is selected through comparison of the confirmed phonetic transcription in step (d) with phonetic transcriptions of stored text.

12. The method of claim 7, wherein the text in step (e) is selected through comparison of the text interpretation with stored text.

13. The method of claim 6, wherein the user confirms the selected text result in step (e).

14. The method of claim 6, wherein the stored confirmed phonetic transcription in step (h) is stored in place of a previously stored phonetic transcription.

15. A method of adapting speech systems, comprising the steps of:

a) receiving a spoken phonebook entry at an automatic speech recognition (ASR) system installed in a vehicle from a user via a vehicle microphone, wherein the ASR system includes a processor configured to execute instructions stored in a non-transitory memory of the ASR system;

b) converting the spoken phonebook entry into a phonetic transcription using the ASR system, wherein the phonetic transcription is a phonetic representation of the user'"'"'s pronunciation of the spoken phonebook entry;

c) presenting the phonetic transcription to the user;

d) receiving confirmation at the vehicle that the phonetic transcription is accurate;

e) selecting text representing the spoken phonebook entry;

f) generating a phonetic interpretation from the selected text in step (e), wherein the phonetic interpretation is a phonetic representation of the selected text;

g) comparing the phonetic interpretation in step f) with the confirmed phonetic transcription of the user'"'"'s pronunciation of the spoken name in step (d); and

h) storing the confirmed phonetic transcription in one or more modules, including memory modules, of the ASR system, a text-to-speech (TTS) system, or both, based on the comparison in step g), wherein the confirmed phonetic transcription represents a user-specific pronunciation of the selected text.

16. The method of claim 15, wherein a text interpretation is generated from the confirmed phonetic transcription in step (d).

17. The method of claim 15, wherein the text in step (e) is selected through comparison of the confirmed phonetic transcription in step (d) with phonetic transcriptions of stored text.

18. The method of claim 16, wherein the text in step (e) is accessed through comparison of the text interpretation with the stored text.

19. The method of claim 15, wherein the user confirms the selected stored text result in step (e).