Information recognition method and apparatus

Patent metadata

Field	Value
Publication number	US-9390711-B2
Application number	US-201414585959-A
Country	US
Kind code	B2
Filing date	Dec 30, 2014
Priority date	Jan 29, 2013
Publication date	Jul 12, 2016
Grant date	Jul 12, 2016

Abstract

An information recognition method and apparatus are provided. The method includes receiving, by a terminal, voice information, extracting a voice feature from the voice information, performing matching calculation on the voice feature and a phoneme string corresponding to each candidate text in multiple candidate texts to obtain a recognition result, where the recognition result includes at least one command word and a label corresponding to the at least one command word, and recognizing, according to the label corresponding to the at least one command word, an operation instruction corresponding to the voice information. A terminal recognizes text information, which is corresponding to voice information input by a user, as an operation instruction.

First claim

What is claimed is:

1. An information recognition method, comprising:
selecting, by a terminal and according to a recognition grammar network, multiple command words from multiple command word-slots to generate multiple candidate texts, wherein the multiple command word-slots comprise multiple command words generated by splitting an action part of a recognition grammar into at least two levels;
receiving, by the terminal, voice information;
converting, by the terminal, the voice information into digital voice information;
extracting, by the terminal, a voice feature from the digital voice information;
performing, by the terminal, a matching calculation on the voice feature and a phoneme string corresponding to each candidate text in the multiple candidate texts to obtain a recognition result, wherein the recognition result comprises at least one command word and a label corresponding to the at least one command word; and
recognizing, by the terminal and according to the label corresponding to the at least one command word, an operation instruction corresponding to the voice information.

2. The information recognition method according to claim 1 , wherein performing the matching calculation comprises:
performing a phoneme distance calculation on the voice feature and the phoneme string corresponding to each candidate text in the multiple candidate texts to obtain a distance value; and
selecting a candidate text, which is corresponding to a phoneme string with a smallest distance value from the voice feature, as the recognition result.

3. The information recognition method according to claim 1 , wherein each command word in the at least one command word is identified by one label, and wherein recognizing the operation instruction comprises recognizing, according to a combination of at least one label corresponding to the at least one command word in the at least one command word, the operation instruction corresponding to the voice information.

4. The information recognition method according to claim 3 , wherein recognizing the operation instruction corresponding to the voice information comprises:
combining a label corresponding to each command word in the at least one command word in the recognition result; and
querying, in a local database or a network server, an operation instruction corresponding to the combination of the label.

5. A terminal comprising:
a network interface;
a processor;
a memory; and
a system bus configured to connect the network interface, the processor, and the memory, wherein the memory is configured to store a computer program, and wherein the processor is configured to run the computer program and cause the terminal to:
select, according to a recognition grammar network, multiple command words from multiple command word-slots to generate multiple candidate texts, wherein the multiple command word-slots comprise multiple command words generated by splitting an action part of a recognition grammar into at least two levels;
receive voice information;
convert the voice information into digital voice information;
extract a voice feature from the digital voice information;
perform a matching calculation on the voice feature and a phoneme string corresponding to each candidate text in the multiple candidate texts to obtain a recognition result, wherein the recognition result comprises at least one command word and a label corresponding to the at least one command word; and
recognize, according to the label corresponding to the at least one command word, an operation instruction corresponding to the voice information.

6. The terminal according to claim 5 , wherein to perform the matching calculation the processor is configured to:
perform phoneme distance calculation on the voice feature and the phoneme string corresponding to each candidate text in the multiple candidate texts to obtain a distance value; and
select a candidate text, which is corresponding to a phoneme string with a smallest distance value from the voice feature, as the recognition result.

7. The terminal according to claim 5 , wherein each command word in the at least one command word is identified by one label, and wherein to recognize the operation instruction corresponding to the voice information the processor is configured to recognize, according to a combination of at least one label corresponding to the at least one command word in the at least one command word, the operation instruction corresponding to the voice information.

8. The terminal according to claim 7 , wherein the processor is configured to:
combine a label corresponding to each command word in the at least one command word in the recognition result; and
query, in a local database or a network server, an operation instruction corresponding to the combination of the label.

9. A computer program product comprising a non-transitory computer readable storage medium storing program code thereon for use by a terminal, the program code comprising instructions for:
selecting, according to a recognition grammar network, multiple command words from multiple command word-slots to generate multiple candidate texts, wherein the multiple command word-slots comprise multiple command words generated by splitting an action part of a recognition grammar into at least two levels;
receiving voice information;
converting the voice information into digital voice information;
extracting a voice feature from the digital voice information;
performing matching calculation on the voice feature and a phoneme string corresponding to each candidate text in the multiple candidate texts to obtain a recognition result, wherein the recognition result comprises at least one command word and a label corresponding to the at least one command word; and
recognizing, according to the label corresponding to the at least one command word, an operation instruction corresponding to the voice information.

Assignees

• Huawei Device Co Ltd

Inventors

• Jiang Hongrui
• Wang Xiyong
• Liang Junbin
• Zheng Weijun
• Zhou Junyang

Classifications

• G10L2015/223

  Execution procedure of a spoken command · CPC title

• G10L2015/088

  Word spotting · CPC title

• G10L15/10

  using distance or distortion measures between unknown speech and reference templates · CPC title

• G10L15/22 Primary

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

• G10L15/19

  Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules · CPC title