MULTILEVEL SPEECH RECOGNITION METHOD AND APPARATUS

• US 20160012820A1
• Filed: 12/02/2014
• Published: 01/14/2016
• Est. Priority Date: 07/09/2014
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: MUN, Minyoung, CHOI, YoungSang
• Granted Patent: US 10,043,520 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

A multilevel speech recognition method and an apparatus performing the method are disclosed. The method includes receiving a first speech command from a user through a speech interface, and extracting a keyword from the first speech command. The method also includes providing a candidate application group of a category providing a service associated with the keyword, and processing a second speech command from the user associated with an application selected from the candidate application group.

Claims

1. A multilevel speech recognition method, the method comprising:

receiving a first speech command from a user through a speech interface;

extracting a keyword from the first speech command;

providing a candidate application group of a category providing a service associated with the keyword; and

processing a second speech command from the user associated with an application selected from the candidate application group.

2. The method of claim 1, wherein the processing comprises:

recognizing the application from the candidate application group using a category speech recognition model, which comprises an acoustic model and a linguistic model based on a set of words associated with the category; and

processing the second speech command associated with the application using an application speech recognition model, which comprises an acoustic model and a linguistic model based on a set of words associated with the application.

3. The method of claim 1, wherein the category is changed based on a use pattern of the user obtained from the first speech command and the second speech command.

4. The method of claim 2, wherein the category speech recognition model is updated based on a characteristic and a use pattern of the user obtained from the first speech command.

5. The method of claim 2, wherein the application speech recognition model is updated based on a characteristic and a use pattern of the user obtained from the second speech command associated with the application.

6. A non-transitory computer-readable storage medium comprising a program comprising instructions to cause a computer to perform the method of claim 1.

7. A multilevel speech recognition method, the method comprising:

recognizing a speech command of a user through a speech interface;

determining, using a processor, whether the speech command comprises an application; and

executing, at the processor, the application in response to the speech command comprising the application.

8. The method of claim 7, wherein the executing further comprises:

processing the speech command of the user associated with the executed application.

9. The method of claim 7, wherein the processing comprises:

processing the speech command of the user associated with the application using an application speech recognition model, which comprises an acoustic model and a linguistic model based on a set of words associated with the application comprised in the speech command.

10. The method of claim 9, wherein the application speech recognition model is updated based on a characteristic and a use pattern of the user obtained from the speech command associated with the application.

11. The method of claim 7, further comprising:

processing a category speech recognition model, comprising an acoustic model and a linguistic model based on words associated with a category, and an application speech recognition model, comprising an acoustic model and a linguistic model based on words associated with an application, to generate a result identifying the speech command to the user terminal.

12. The method of claim 11, further comprising:

continually updating the category speech recognition model and the application speech recognition model.

13. The method of claim 7, wherein the recognizing comprises

recognizing a word frequently received through a microphone, and

recognizing the frequently received word associated with an application.

14. The method of claim 7, wherein the recognizing comprises

analyzing a use pattern and a voice of the user based on a word included in the speech command and adaptively updating the determination based on a result of the analyzing.

15. A non-transitory computer-readable storage medium comprising a program comprising instructions to cause a computer to perform the method of claim 7.

16. A user terminal, comprising:

a speech interface; and

a processor operatively connected to the speech interface,

wherein the processor is configured to receive a first speech command of a user through the speech interface, extract a keyword from the first speech command, provide a candidate application group of a category providing a service associated with the keyword, and process a second speech command from the user associated with an application selected from the candidate application group.

17. The user terminal of claim 16, wherein the processor is further configured to recognize the application selected using an acoustic model and a linguistic model based on a set of words associated with the category, and

process the second speech command associated with the application using an acoustic model and a linguistic model based on a set of words associated with the application.

18. The user terminal of claim 16, wherein the processor is further configured to recognize the application from the candidate application group using a category speech recognition model, which comprises an acoustic model and a linguistic model based on a set of words associated with the category, and process the second speech command associated with the application using an application speech recognition model, which comprises an acoustic model and a linguistic model based on a set of words associated with the application.

19. The user terminal of claim 18, wherein the processor is further configured to continually update the category speech recognition model and the application speech recognition model.

20. The user terminal of claim 18, wherein the processor is further configured to analyze a use pattern and a voice of the user based on a word included in the first speech command and adaptively update the category speech recognition model and the application speech recognition model based on a result of the analyzing.