METHOD AND APPARATUS FOR INITIATING AN OPERATION USING VOICE DATA

• US 20170110125A1
• Filed: 10/13/2016
• Published: 04/20/2017
• Est. Priority Date: 10/14/2015
• Status: Abandoned Application

• Current Assignee: Alibaba Group Holding Ltd.
• Original Assignee: Alibaba Group Holding Ltd.
• Inventors: XU, Minqiang, YAN, Zhijie, GAO, Jie, CHU, Min
• Application Number:
  US15/292,632
• Publication Number:
  US 20170110125A1
• Time in Patent Office: Days
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/04 Segmentation; Word boundary detection
• 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 17/04 Training, enrolment or model building
• CPC Class Codes: G10L 17/08 Use of distortion metrics or a particular distance between probe pattern and reference templates
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10

Abstract

A method for initiating an operation using voice is provided. The method includes extracting one or more voice features based on first audio data detected in a use stage; determining a similarity between the first audio data and a preset first voice model according to the one or more voice features, wherein the first voice model is associated with second audio data of a user, and the second audio data is associated with one or more preselected voice contents; and executing an operation corresponding to the first voice model based on the similarity.

Claims

1. A method for initiating an operation using voice, comprising:

extracting one or more voice features based on first audio data ;

determining a similarity between the first audio data and a preset first voice model according to the one or more voice features, wherein the first voice model is associated with second audio data of a user, and the second audio data is associated with one or more preselected voice contents; and

executing an operation corresponding to the first voice model based on the similarity.

2. The method according to claim 1, wherein the step of extracting one or more voice features comprises:

determining whether the first audio data is voice data after the first audio data is detected in the use stage;

if the first audio data is voice data, extracting the one or more voice features based on the first audio data; and

if the first audio data is not voice data, discarding the first audio data.

3. The method according to claim 1, wherein the step of extracting one or more voice features comprises:

segmenting the first audio data into one or more pieces of voice segment data, wherein each piece of the voice segment data is associated with a voice content; and

extracting one or more voice features of each piece of the voice segment data.

4. The method according to claim 3, wherein the preset first voice model comprises one or more voice sub-models, and each voice sub-model is associated with audio data of a predetermined voice content of the user, and wherein the step of determining the similarity between the first audio data and a preset first voice model comprises:

identifying a voice sub-model corresponding to each piece of the voice segment data according to a segmenting order;

determining the voice segment similarity between the one or more voice features of each piece of the voice segment data and the voice sub-model; and

determining the similarity between the first audio data and the first voice model according to each voice segment similarity.

5. The method according to claim 1, wherein the step of executing an operation corresponding to the voice model comprises:

executing the operation corresponding to the first voice model when the similarity is greater than a preset similarity threshold, and

wherein a screen of a device is in a screen-lock state in the use stage, and the operation corresponding to the first voice model comprises an unlock operation and a starting of an application.

6. The method according to claim 1, further comprising:

obtaining one or more pieces of audio data of the user in a registration stage;

training a second voice model according to the one or more pieces of the audio data, wherein the one or more pieces of the audio data is associated with one or more voice contents of the user, and the one or more voice contents are different from the one or more preselected voice contents; and

training the first voice model according the one or more pieces of the audio data and the second voice model.

7. The method according to claim 6, wherein the step of obtaining one or more pieces of audio data of the user in a registration stage comprises:

determining whether a piece of the audio data is voice data after the piece of audio data is detected in the registration stage;

if the piece of audio data is voice data, determining that the piece of audio data is associated with the user; and

if the piece of audio data is not voice data, discarding the piece of audio data.

8. The method according to claim 6, wherein the step of training a second voice model according to the one or more pieces of the audio data comprises:

identifying a preset third voice model, wherein the third voice model is associated with audio data of one or more speakers different from the user, and the audio data of one or more speakers is associated with at least one voice content different from each of the one or more preselected voice contents; and

training the second voice model by using the one or more pieces of the audio data and the third voice model.

9. The method according to claim 6, wherein the first voice model comprises one or more voice sub-models, and wherein the step of training the first voice model comprises:

segmenting each piece of the audio data of the user into one or more pieces of voice segment data, wherein each piece of the voice segment data is associated with a voice content;

extracting at least one voice feature from each piece of the voice segment data; and

training the first voice model by using the at least one voice feature of each piece of the voice segment data and the second voice model.

10. The method according to claim 6, further comprising:

updating the first voice model and the second voice model based on the first audio data detected in the use stage.

11. An apparatus for initiating an operation using voice, comprising:

a voice feature extracting module configured to extract one or more voice features based on first audio data;

a model similarity determining module configured to determine a similarity between the first audio data and a preset first voice model according to the one or more voice features, wherein the first voice model is associated with second audio data of a user, and the second audio data is associated with one or more preselected voice contents; and

an operation executing module configured to execute an operation corresponding to the first voice model based on the similarity.

12. The apparatus according to claim 11, wherein the voice feature extracting module comprises:

a first voice data determining sub-module configured to determine whether the first audio data is voice data, invoking an extracting sub-module, and if not, invoking a first discarding sub-module;

a first extracting sub-module configured to extract the one or more voice features based on the audio data, wherein the first extracting sub-module is invoked if the first voice data determining sub-module determines that the first audio data is voice data; and

a first discarding sub-module configured to discard the audio data, wherein the first discarding sub-module is invoked if the first voice data determining sub-module determines that the first audio data is not voice data.

13. The apparatus according to claim 11, wherein the voice feature extracting module comprises:

a first segmenting sub-module configured to segment the first audio data into one or more pieces of voice segment data, wherein each piece of the voice segment data is associated with a voice content; and

a second extracting sub-module configured to extract one or more voice features of each piece of the voice segment data.

14. The apparatus according to claim 13, wherein the preset first voice model comprises one or more voice sub-models, and each voice sub-model is associated with audio data of a predetermined voice content of the user, and wherein the model similarity determining module comprises:

a voice sub-model identifying sub-module configured to identify a voice sub-model corresponding to each piece of the voice segment data according to a segmenting order;

a voice segment similarity determining sub-module configured to determine the voice segment similarity between the one or more voice features of each piece of the voice segment data and the voice sub-model; and

a similarity determining sub-module configured to determine the similarity between the first audio data and the first voice model according to each voice segment similarity.

15. The apparatus according to claim 11, wherein the operation executing module comprises:

an executing sub-module configured to execute the operation corresponding to the first voice model when the similarity is greater than a preset similarity threshold, and

wherein a screen of a device is in a screen-lock state in the use stage, and the operation corresponding to the first voice model comprises an unlock operation and a starting of an application.

16. The apparatus according to claim 11, further comprising:

an audio data obtaining module configured to obtain one or more pieces of audio data of the user in a registration stage;

a second voice model training module configured to train a second voice model according to the one or more pieces of the audio data, wherein the one or more pieces of the audio data is associated with one or more voice contents of the user, and the one or more voice contents are different from the one or more preselected voice contents; and

a first voice model training module configured to train the first voice model according the one or more pieces of the audio data and the second voice model.

17. The apparatus according to claim 16, wherein the audio data obtaining module comprises:

a second voice data determining sub-module configured to determine whether a piece of the audio data is voice data after the piece of audio data is detected in the registration stage;

a determining sub-module configured to determine that the piece of audio data is associated with the user, wherein the determining sub-module is invoked if the second voice data determining sub-module determines that the piece of audio data is voice data; and

a second discarding sub-module configured to discard the piece of audio data, wherein the second discarding sub-module is invoked if the second voice data determining sub-module determines that the piece of audio data is not voice data.

18. The apparatus according to claim 16, wherein the second voice model training module comprises:

a third voice model identifying sub-module configured to identify a preset third voice model, wherein the third voice model is associated with audio data of one or more speakers different from the user, and the audio data of one or more speakers is associated with at least one voice content different from each of the one or more preselected voice contents; and

a first training sub-module configured to train the second voice model by using the one or more pieces of the audio data and the third voice model.

19. The apparatus according to claim 16, wherein the first voice model comprises one or more voice sub-models, and wherein the first voice model training module comprises:

a second segmenting sub-module configured to segment each piece of the audio data of the user into one or more pieces of voice segment data, wherein each piece of the voice segment data is associated with a voice content;

a third extracting sub-module configured to extract at least one voice feature from each piece of the voice segment data; and

a second training sub-module configured to train the first voice model by using the at least one voice feature of each piece of the voice segment data and the second voice model.

20. The apparatus according to claim 16, further comprising:

a model updating module configured to update the first voice model and the second voice model based on the first audio data detected in the use stage.

21. A non-transitory computer readable medium that stores a set of instructions that is executable by at least one processor of an electronic device to cause the electronic device to perform a method for initiating an operation using voice, the method comprising:

extracting one or more voice features based on first audio data;

determining a similarity between the first audio data and a preset first voice model according to the one or more voice features, wherein the first voice model is associated with second audio data of a user, and the second audio data is associated with one or more preselected voice contents; and

executing an operation corresponding to the first voice model based on the similarity.

22. The non-transitory computer readable medium of claim 21, wherein the set of instructions that is executable by the at least one processor of the electronic device to cause the electronic device to further perform:

determining whether the first audio data is voice data;

if the first audio data is voice data, extracting the one or more voice features based on the first audio data; and

if the first audio data is not voice data, discarding the first audio data.

23. The non-transitory computer readable medium of claim 21, wherein the set of instructions that is executable by the at least one processor of the electronic device to cause the electronic device to further perform:

segmenting the first audio data into one or more pieces of voice segment data, wherein each piece of the voice segment data is associated with a voice content; and

extracting one or more voice features of each piece of the voice segment data.

24. The non-transitory computer readable medium of claim 23, the preset first voice model comprises one or more voice sub-models, and each voice sub-model is associated with audio data of a predetermined voice content of the user, and wherein the set of instructions that is executable by the at least one processor of the electronic device to cause the electronic device to further perform:

identifying a voice sub-model corresponding to each piece of the voice segment data according to a segmenting order;

determining the voice segment similarity between the one or more voice features of each piece of the voice segment data and the voice sub-model; and

determining the similarity between the first audio data and the first voice model according to each voice segment similarity.

25. The non-transitory computer readable medium of claim 21, wherein the set of instructions that is executable by the at least one processor of the electronic device to cause the electronic device to further perform:

obtaining one or more pieces of audio data of the user in a registration stage;

training a second voice model according to the one or more pieces of the audio data, wherein the one or more pieces of the audio data is associated with one or more voice contents of the user, and the one or more voice contents are different from the one or more preselected voice contents; and

training the first voice model according the one or more pieces of the audio data and the second voice model.

26. The non-transitory computer readable medium of claim 25, wherein the set of instructions that is executable by the at least one processor of the electronic device to cause the electronic device to further perform:

identifying a preset third voice model, wherein the third voice model is associated with audio data of one or more speakers different from the user, and the audio data of one or more speakers is associated with at least one voice content different from each of the one or more preselected voice contents; and

training the second voice model by using the one or more pieces of the audio data and the third voice model.