Speech recognition method, speech wakeup apparatus, speech recognition apparatus, and terminal

Patent metadata

Field	Value
Publication number	US-10943584-B2
Application number	US-201715729097-A
Country	US
Kind code	B2
Filing date	Oct 10, 2017
Priority date	Apr 10, 2015
Publication date	Mar 9, 2021
Grant date	Mar 9, 2021

Abstract

Embodiments of the present invention provide a speech recognition method and a terminal. The method includes: listening, by a speech wakeup apparatus, to speech information in a surrounding environment; when determining that the speech information obtained by listening matches a speech wakeup model, buffering, by the speech wakeup apparatus, speech information, of first preset duration, obtained by listening, and sending a trigger signal for triggering enabling of a speech recognition apparatus, where the trigger signal is used to instruct the speech recognition apparatus to read and recognize the speech information buffered by the speech wakeup apparatus; and recognizing first speech information buffered by the speech wakeup apparatus and the second speech information obtained by listening, to obtain a recognition result.

First claim

What is claimed is:

1. A speech processing method, wherein the method is applied in a terminal comprising a speech wakeup apparatus and a speech recognition apparatus, the method comprising:
listening, by the speech wakeup apparatus, to a first speech information, wherein the first speech information comprises a wakeup information for enabling the speech recognition apparatus and a first portion of a command word;
determining that the wakeup information matches a speech wakeup model, and enabling the speech recognition apparatus;
listening, by the speech recognition apparatus, to a second speech information after being enabled, wherein the second speech information comprises a second portion of the command word; and
obtaining, by the speech recognition apparatus, a speech instruction information according to the first speech information and the second speech information, wherein the speech instruction information matches the command word, and the command word comprises the first portion of the command word and the second portion of the command word.

2. The method according to claim 1 , further comprising performing an operation according to the speech instruction information.

3. The method according to claim 2 , wherein the performing an operation according to the speech instruction information comprises: when the speech instruction information matches a pre-set speech instruction information, performing the operation according to the pre-set speech instruction information.

4. The method according to claim 1 , further comprising: generating, by the speech wakeup apparatus, a trigger signal for enabling the speech recognition apparatus when the wakeup information matches a speech wakeup model, and sending the trigger signal to the speech recognition apparatus.

5. The method according to claim 1 , wherein the determining that the wakeup information matches a speech wakeup model further comprises: if the wakeup information matches a predetermined wakeup speech information, extracting a voiceprint feature from the wakeup information, and determining the extracted voiceprint feature matches a predetermined voiceprint feature.

6. The method according to claim 5 , wherein the voiceprint feature comprises an acoustic parameter that reflects the voiceprint feature, wherein the acoustic parameter comprises a pitch contour, a linear prediction coefficient, a spectral envelope parameter, a harmonic energy ratio, a resonant peak frequency and its bandwidth, a cepstrum, or a Mel-frequency cepstrum coefficient.

7. The method according to claim 1 , wherein before the listening to a first speech information: the speech wakeup apparatus is in a state for listening to a speech information in a surrounding environment, and the speech recognition apparatus is inactive.

8. The method according to claim 7 , wherein the listening to the speech information in the surrounding environment is executed in: a standby state, a non-standby state, or a screen-locked state.

9. The method according to claim 8 , further comprising pre-storing pre-set speech instruction information.

10. The method according to claim 1 , wherein the speech wakeup apparatus is a digital signal processor, and the speech recognition apparatus is an application processor or a CPU.

11. The method according to claim 1 , wherein the listening to a first speech information comprises listening to the first speech information within a first preset duration, and the listening to a second speech information comprises listening to the second speech information within a second preset duration.

12. The method according to claim 11 , further comprising disabling speech recognition function of the speech recognition apparatus automatically when a further trigger signal is not received again within a third preset duration after a previous trigger signal is received.

13. A non-transitory computer-readable medium having computer-executable instructions stored thereon for execution by a processor, wherein the instructions cause the processor to execute the method according to claim 1 .

14. A terminal, comprising a speech wakeup apparatus and a speech recognition apparatus,
the speech wakeup apparatus configured to listen to a first speech information, wherein the first speech information comprises a wakeup information for enabling the speech recognition apparatus and a first portion of a command word;
the speech recognition apparatus configured to listen to a second speech information after being enabled, wherein the second speech information comprises a second portion of the command word; and
the speech recognition apparatus configured to obtain a speech instruction information according to the first speech information and the second speech information, wherein the speech instruction information matches the command word, and the command word comprises the first portion of the command word and the second portion of the command word.

15. The terminal according to claim 14 , further comprising an execution module configured to perform an operation according to the speech instruction information.

16. The terminal according to claim 14 , the speech wakeup apparatus further configured to: generate a trigger signal for enabling the speech recognition apparatus when the wakeup information matches a speech wakeup model, and send the trigger signal to the speech recognition apparatus; and
the speech recognition apparatus further configured to enable speech recognition function according to the trigger signal.

17. The terminal according to claim 14 , the speech wakeup apparatus further configured to: extract a voiceprint feature from the wakeup information if the wakeup information matches a predetermined wakeup speech information, and determine the extracted voiceprint feature matches a predetermined voiceprint feature.

18. The terminal according to claim 14 , wherein before listening to the first speech information: the speech wakeup apparatus is in a state for listening to a speech information in a surrounding environment, and the speech recognition apparatus is inactive.

19. The terminal according to claim 14 , wherein before listening to the first speech information, the terminal is in: a standby state, a non-standby state, or a screen-locked state.

20. The terminal according to claim 14 , the speech wakeup apparatus further configured to listen to the first speech information within a first preset duration; and
the speech recognition apparatus is further configured to listen to the second speech information within a second preset duration.

21. The terminal according to claim 20 , the speech recognition apparatus further configured to disable a speech recognition function of the speech recognition apparatus automatically when a further trigger signal is not received again within a third preset duration after the previous trigger signal is received.

Assignees

• Huawei Tech Co Ltd

Inventors

• Zhou Junyang

Classifications

• G10L2015/223

  Execution procedure of a spoken command · CPC title

• G10L15/00

  Speech recognition (G10L17/00 takes precedence) · CPC title

• H04M1/725

  Cordless telephones (user interfaces specially adapted therefor H04M1/724) · CPC title

• G10L15/02

  Feature extraction for speech recognition; Selection of recognition unit · CPC title

• G10L15/22 Primary

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title