Method, device and system for implementing speech recognition function

• US 8,417,521 B2
• Filed: 04/18/2008
• Issued: 04/09/2013
• Est. Priority Date: 10/21/2005
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: Chen, Cheng
• Primary Examiner(s): Han, Qi
• Application Number:
  US12/105,620
• Publication Number:
  US 20080228483A1
• Time in Patent Office: 1,817 Days
• Field of Search: 704/231, 704/235, 704/251, 704/270, 704/270.1
• US Class Current: 704/231
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: H04L 65/1101 Session protocols
• CPC Class Codes: H04L 65/1106 Call signalling protocols; H.323 and related
• CPC Class Codes: H04L 65/762 at the source reformatting of additional data in video distribution servers H04N21/2355

Abstract

The present disclosure discloses a method, a device and a system for implementing a speech recognition function, in which a media resource control device controls a media resource processing device to recognize a speech input by a user via H.248 protocol. The method includes receiving, by the media resource processing device, an H.248 message carrying a speech recognition instruction and a related parameter sent by the media resource control device; performing speech recognition according to the speech recognition instruction and the parameter; and reporting a recognition result to the media resource control device. A corresponding device and system for implementing the speech recognition function is further provided.

Claims

1. A method for recognizing a speech, comprising:

receiving, by a media resource processing device, an H.248 message carrying a speech recognition instruction and a related parameter sent by a media resource control device, wherein the related parameter comprises a recognition grammar parameter for the speech, the recognition grammar parameter for the speech is a recognition grammar for a common speech;

performing, by the media resource processing device, speech recognition according to the speech recognition instruction and the related parameter; and

reporting a recognition result to the media resource control device,

the method further comprise;

extracting, by the media resource processing device, a text string and performing the speech recognition after receiving the recognition grammar parameter, wherein a specific character string is embedded into the recognition grammar parameter.

2. The method according to claim 1, wherein when the recognition grammar for the speech is prestored on the media resource processing device or an external server, the recognition grammar parameter is an ID and storage location information of a recognition grammar file, and the media resource processing device reads and caches the recognition grammar file according to the storage location information and performs the speech recognition according to the recognition grammar file which is read.

3. The method according to claim 1, during the speech recognition performed by the media resource processing device, further comprising at least one of:

detecting, by the media resource processing device, an abnormal event comprising the waiting time for a user to input timing out, a recognition timeout, a mismatch between a user input and grammar, the grammar file not existing, the grammar file being read error, a user input being unable to be recognized, a recognition error, and a software/hardware error of the media resource processing device; and

detecting, by the media resource processing device, an event of inputting a voice input by a user and reporting a type of a detected voice input by the user to the media resource control device, the voice input comprising;

a dual tone multi-frequency (DTMF) and a common speech.

4. The method according to claim 3, further comprising detecting, by the media resource processing device, a speech recognition process according to an instruction of the media resource control device, and feeding back a detection result to the media resource control device.

5. The method according to claim 4, further comprising feeding back, by the media resource processing device, a corresponding error code to the media resource control device when detecting the abnormal event during the speech recognition.

6. The method according to claim 1, wherein the specific character includes one or more grammar rules of speech recognition, the method further comprising:

extracting, by the media resource processing device, the one or more grammar rules after receiving the specific character; and

performing the speech recognition according to the one or more grammar rules.

7. The method according to claim 6, wherein at least one of the grammar rules define permissible results of the speech recognition.