Method for recognizing voice signal and electronic device supporting the same

• US 10,354,643 B2
• Filed: 10/13/2016
• Issued: 07/16/2019
• Est. Priority Date: 10/15/2015
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Chakladar, Subhojit, Kim, Junhui
• Primary Examiner(s): Colucci, Michael C
• Application Number:
  US15/292,951
• Publication Number:
  US 20170110117A1
• Time in Patent Office: 1,006 Days
• Field of Search: 704 9, 704255, 704254, 704243, 704240, 704235, 704234, 706 12, 703 2, 700 83
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/142 Hidden Markov Models [HMMs]
• CPC Class Codes: G10L 15/16 using artificial neural networks
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

An electronic device is provided including at least one microphone, a communication circuit, a processor and a memory, wherein the memory stores at least one application program or a software program executing a voice instruction, which is triggered in response to a voice input, upon the performance, the memory stores instructions to allow the processor to sequentially receive a plurality of utterances including a first speech element from a first user through the at least one microphone, generate a voice recognition model of the first user on the basis of at least some of the plurality of utterances, store the generated voice recognition model in the memory, and transmit the generated voice recognition model of the first user to the outside through the communication circuit so that a first external device uses the generated voice recognition model of the first user.

Claims

1. An electronic device, comprising:

at least one microphone;

a communication circuit;

a processor electrically connected to the at least one microphone and the communication circuit; and

a memory electrically connected to the processor,

wherein the memory;

stores at least one of an application program and a software program executing a voice instruction, which is triggered in response to a voice input; and

stores instructions that cause, when executed by the processor, the processor to;

sequentially receive a plurality of utterances including a first speech element from a first user through the at least one microphone,

generate a voice recognition model of the first user on the basis of at least some of the plurality of utterances, the voice recognition model comprising at least one of characteristic information of the at least one microphone and environment information surrounding the electronic device,

store the generated voice recognition model in the memory,

transmit the generated voice recognition model of the first user to the outside through the communication circuit so that a first external device uses the generated voice recognition model of the first user when performing voice recognition on the utterances including the first speech element of the first user,

calculate an error rate occurring when at least one of the application program and the software program is triggered on the basis of at least one parameter generated on the basis of at least one of the plurality of utterances,

determine, based on at least one of the characteristic information of the at least one microphone and the environment information, whether the calculated error rate is greater than or equal to a threshold value,

when the calculated error rate is greater than or equal to the threshold value, change the at least one parameter, count a repetition frequency of the changed at least one parameter, and if the counted repetition frequency reaches the selected frequency, associate the changed at least one parameter with a voice signal, and

when the calculated error rate is less than the threshold value, associate the at least one parameter with the voice signal.

2. The electronic device of claim 1, wherein the instructions cause the processor, when executed by the processor, to:

store the plurality of received utterances in the memory, and

transmit data or signals including or representing at least some of the plurality of stored utterances to the outside through the communication circuit.

3. The electronic device of claim 1, wherein the instructions cause the processor, when executed by the processor, to transmit the generated voice recognition model of the first user to a server, which communicates with the first electronic device, through the communication circuit.

4. The electronic device of claim 1, wherein after the voice recognition model of the first user is generated, the instructions cause the processor, when executed by the processor, to:

receive the utterances including the first speech element from the first user through the microphone,

perform the voice recognition on the received utterances on the basis of the voice recognition model of the first user, and

execute at least one of the application program and the software program on the basis of a result of the voice recognition.

5. The electronic device of claim 1, wherein the voice recognition model is associated with at least one of a hidden Markov model (HMM) and an artificial neural network (ANN).

6. The electronic device of claim 1, wherein the instructions cause the processor, when executed by the processor, to calculate the error rate on the basis of the at least one changed parameter when the calculated error rate is equal to or greater than the threshold value.

7. The electronic device of claim 6, wherein the instructions cause the processor, when executed by the processor, to repeat a change in the at least one parameter and the calculation of the error rate until the calculated error rate is less than the threshold value when the calculated error rate is greater than or equal to the threshold value.

8. An electronic device, comprising:

a communication circuit configured to communicate with a first external device including a first microphone and a second external electronic device including a second microphone;

a processor electrically connected to the communication circuit; and

a memory electrically connected to the processor,

wherein the memory stores instructions that cause, when executed by the processor, the processor to;

receive a voice recognition model of the first user, which is generated by sequentially receiving a plurality of utterances including a first speech element received from a first user through the first microphone of the first external electronic device from the first external electronic device through the communication circuit, the voice recognition model comprising at least one of characteristic information of the at least one microphone and environment information surrounding the electronic device,

store the received voice recognition model of the first user in the memory,

receive a transmission request of the voice recognition model of the first user from the second external electronic device through the communication circuit, and

transmit the generated voice recognition model of the first user to the second external electronic device through the communication circuit so that the second external electronic device uses the generated voice recognition model of the first user when performing the voice recognition on the utterances including the first speech element of the first user in response to the received request,

wherein at least one of the characteristic information of the at least one microphone and the environment information is used by at least one of the first external electronic device and the second external electronic device to determine whether an error rate is greater than or equal to a threshold value,

wherein the error rate occurs when at least one of the first external electronic device and the second external electronic device performs voice recognition on the basis of at least one parameter generated on the basis of at least one of the plurality of utterances, and

wherein, when the calculated error rate is greater than or equal to the threshold value, the at least one parameter is changed, a repetition frequency of the changed at least one parameter is counted, and if the counted repetition frequency reaches a selected frequency, the changed at least one parameter is associated with a voice signal, and

wherein, when the calculated error rate is less than the threshold value, the at least one parameter is associated with the voice signal.

9. The electronic device of claim 8, wherein the voice recognition model is associated with at least one of a hidden Markov model (HMM) and an artificial neural network (ANN).

10. An electronic device, comprising:

a communication circuit;

at least one microphone;

a processor electrically connected to the communication circuit and the at least one microphone; and

a memory electrically connected to the processor,

wherein the memory stores at least one of an application program and a software program executing voice instructions, which is triggered in response to a voice input, and stores instructions that cause, when executed by the processor, the processor to;

transmit a signal requesting a voice recognition model of a first user, which is generated by sequentially receiving a plurality of utterances including a first speech element received from the first user through a first microphone of a first external electronic device to the outside through the communication circuit,

receive the voice recognition model of the first user provided as a response to the requested signal through the communication circuit, the voice recognition model comprising at least one of characteristic information of the at least one microphone and environment information surrounding the electronic device,

store the received voice recognition model of the first user in the memory,

receive the utterances including the first speech element from the first user through the at least one microphone,

perform voice recognition on the received utterances on the basis of at least some of the stored voice recognition model of the first user,

calculate an error rate occurring when at least one of the application program and the software program is triggered on the basis of at least one parameter generated on the basis of at least one of the plurality of utterances,

determine, based on at least one of the characteristic information of the at least one microphone and the environment information, whether the calculated error rate is greater than or equal to a threshold value,

when the calculated error rate is greater than or equal to the threshold value, changing the at least one parameter, counting a repetition frequency of the changed at least one parameter, and if the counted repetition frequency reaches the selected frequency, associating the changed at least one parameter with a voice signal, and

when the calculated error rate is less than the threshold value, associating the at least one parameter with the voice signal.

11. The electronic device of claim 10, wherein the instructions cause the processor, when executed by the processor, to at least partially change the received voice recognition model of the first user on the basis of at least some of the utterances including the first speech element of the first user received through the at least one microphone.

12. The electronic device of claim 11, wherein after the received voice recognition model of the first user is at least partially changed, the instructions cause the processor, which when executed by the processor, to:

receive the utterances including the first speech element from the first user through the at least one microphone, and

perform the voice recognition on the received utterances on the basis of at least some of the at least partially changed voice recognition model of the first user.

13. The electronic device of claim 11, wherein the voice recognition model is associated with at least one of a hidden Markov model (HMM) and an artificial neural network (ANN).

14. A voice recognition method of an electronic device, comprising:

sequentially receiving a plurality of utterances including a first speech element from a first user though at least one microphone;

generating a voice recognition model of the first user on the basis of at least some of the plurality of utterances, the voice recognition model comprising at least one of characteristic information of the at least one microphone and environment information surrounding the electronic device;

storing the generated voice recognition model in the memory;

transmitting the generated voice recognition model of the first user to the outside through the communication circuit so that a first external device uses the generated voice recognition model of the first user when performing voice recognition on the utterances including the first speech element of the first user;

calculating an error rate occurring when at least one of the application program and the software program is triggered;

determining, based on at least one of the characteristic information of the at least one microphone and the environment information, whether the calculated error rate is greater than or equal to a threshold value; and

when the calculated error rate is greater than or equal to the threshold value, changing the at least one parameter, counting a repetition frequency of the changed at least one parameter, and if the counted repetition frequency reaches the selected frequency, associating the changed at least one parameter with the voice signal, and

associating the at least one parameter with the voice signal when the calculated error rate is less than the threshold value.

15. The voice recognition method of claim 14, further comprising:

storing the plurality of received utterances in the memory; and

transmitting data or signals including or representing at least some of the plurality of stored utterances to the outside through the communication circuit.

16. The voice recognition method of claim 14, wherein transmitting the generated first voice recognition model of the first user to the outside through the communication circuit includes transmitting the generated voice recognition model of the first user to a server, which communicates with the first electronic device, through the communication circuit.

17. The voice recognition method of claim 14, wherein after the voice recognition model of the first user is generated, receiving the utterances including the first speech element from the first user through the microphone includes performing the voice recognition on the received utterances on the basis of the voice recognition model of the first user and executing at least one of the application program and the software program on the basis of the result of the voice recognition.

18. The voice recognition method of claim 14, wherein the voice recognition model is associated with at least one of a hidden Markov model (HMM) and an artificial neural network (ANN).

19. The voice recognition method of claim 14, further comprising:

calculating the error rate on the basis of the at least one changed parameter when the calculated error rate is greater than or equal to the threshold value.

20. The voice recognition method of claim 19, further comprising:

repeating a change in the at least one parameter and a calculation of the error rate until the calculated error rate is less than the threshold value when the calculated error rate is equal to or greater than the threshold value.

21. A voice recognition method of an electronic device, comprising:

receiving a voice recognition model of a first user, which is generated by sequentially receiving a plurality of utterances including a first speech element received from the first user through a first microphone of a first external electronic device from the first external electronic device, through a communication circuit, the voice recognition model comprising at least one of characteristic information of the at least one microphone and environment information surrounding the electronic device;

storing the received voice recognition model of the first user in a memory;

receiving a transmission request of the voice recognition model of the first user from a second external electronic device through the communication circuit; and

transmitting the generated voice recognition model of the first user to the second external electronic device through the communication circuit so that the second external electronic device uses the generated voice recognition model of the first user when performing the voice recognition on the utterances including the first speech element of the first user in response to the received request,

wherein at least one of the characteristic information of the at least one microphone and the environment information is used by at least one of the first external electronic device and the second external electronic device to determine whether an error rate is greater than or equal to a threshold value,

wherein the error rate occurs when at least one of the first external electronic device and the second external electronic device performs voice recognition on the basis of at least one parameter generated on the basis of at least one of the plurality of utterances,

wherein, when the calculated error rate is greater than or equal to the threshold value, the at least one parameter is changed, a repetition frequency of the changed at least one parameter is counted, and if the counted repetition frequency reaches a selected frequency, the changed at least one parameter is associated with a voice signal, and

wherein, when the calculated error rate is less than the threshold value, the at least one parameter is associated with the voice signal.

22. The voice recognition method of claim 21, wherein the voice recognition model is associated with at least one of a hidden Markov model (HMM) and an artificial neural network (ANN).

23. A voice recognition method of an electronic device, comprising:

transmitting a signal requesting a voice recognition model of a first user, which is generated by sequentially receiving a plurality of utterances including a first speech element received from the first user through a first microphone of a first external electronic device, to the outside through a communication circuit;

receiving the voice recognition model of the first user provided as a response to the requested signal through the communication circuit, the voice recognition model comprising at least one of characteristic information of the at least one microphone and environment information surrounding the electronic device;

storing the received voice recognition model of the first user in a memory;

receiving the utterances including the first speech element from the first user through the at least one microphone;

performing the voice recognition on the received utterances on the basis of at least some of the stored voice recognition model of the first user;

calculating an error rate occurring when at least one of the application program and the software program is triggered on the basis of at least one parameter generated on the basis of at least one of the plurality of utterances;

determining, based on at least one of the characteristic information of the at least one microphone and the environment information, whether the calculated error rate is greater than or equal to a threshold value; and

changing, when the calculated error rate is greater than or equal to the threshold value, the at least one parameter, counting a repetition frequency of the changed at least one parameter, and if the counted repetition frequency reaches the selected frequency, associating the changed at least one parameter with the voice signal, and

when the calculated error rate is less than the threshold value, associating the at least one parameter with the voice signal.

24. The voice recognition method of claim 23, further comprising:

at least partially changing the received voice recognition model of the first user on the basis of at least some of the utterances including the first speech element of the first user received through the at least one microphone.

25. The voice recognition method of claim 24, further comprising:

after the received voice recognition model of the first user is at least partially changed, receiving the utterances including the first speech element from the first user through the at least one microphone; and

performing the voice recognition on the received utterances on the basis of the at least partially changed voice recognition model of the first user.

26. The voice recognition method of claim 24, wherein the voice recognition model is associated with at least one of a hidden Markov model (HMM) and an artificial neural network (ANN).