Voice recognition system and construction method thereof

• US 9,953,639 B2
• Filed: 06/26/2014
• Issued: 04/24/2018
• Est. Priority Date: 08/04/2013
• Status: Active Grant

• Current Assignee: Hua Xu
• Original Assignee: Hua Xu
• Inventors: Xu, Hua
• Primary Examiner(s): Pullias, Jesse
• Application Number:
  US14/909,741
• Publication Number:
  US 20160171978A1
• Time in Patent Office: 1,398 Days
• Field of Search: 704 9, 704 10, 704231-257, 704270-275
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• 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

Abstract

Disclosed are a voice recognition system and a construction method for the voice recognition system. By way of layering the system, a general semantic recognition operation for the system is separated from a specific semantic recognition operation for an application program; and by way of classifying the application programs and abstracting out a common performance function, the system can find the application program matching the voice content semantics very efficiently and a third-party program is easily added into the existing voice recognition system. The present invention maps the performance function to a regular expression with semantic variables, so that the system can recognize more semantic expression manners with optimization of the semantic recognition. Therefore, the system can show more humanized characteristics.

Claims

1. A construction method based on a voice recognition system, the voice recognition system comprising a voice recognition module and a registration centre for storing information of application programs installed on the voice recognition system and establishing a multi fork tree structure of the voice recognition system, the registration centre comprising a root node for storing a plurality of regular expressions with semantic variables to find and open files or the application programs, a plurality of class nodes each for storing a plurality of regular expressions with semantic variables which are used to describe common behaviors of a series of the application programs and mapping the regular expressions to a public function set;

a plurality of application program nodes each for representing one of the application programs that belongs to one of the class nodes and inherits the public function set of the class node, and each application program node comprises a plurality of regular expressions with semantic variables that are not used in the class nodes and mapped to a specific function call set;

and a plurality of application program internal nodes for storing a plurality of regular expressions used in the application programs and corresponding function sets, the method comprising;

step S01;

the application programs submitting information to the registration centre;

step S02;

the voice recognition system receiving a voice input by a user, and the voice recognition module converting voice content into text form;

Step S03;

the voice recognition system determining whether the voice content matches with the regular expressions with semantic variables of the root node or not, wherein if the voice content matches with the regular expressions with semantic variables of the root node, a first function is called; and

wherein if the voice content does not match with the regular expressions with semantic variables of the root node, a next step is performed;

step S04;

the class nodes receiving the voice content, wherein if the voice content does not match with the regular expressions with semantic variables of any class node, step S05 is performed;

and wherein if the voice content matches with the regular expressions with semantic variables of one class node, step S07 is performed;

step S05;

the voice content being delivered to all application nodes wherein if the voice content matches with the regular expressions with semantic variables of one of application program nodes, step S07 is performed; and

wherein if the voice content does not match with the regular expressions with semantic variables of any application program nodes, step S06 is performed;

step S06;

the voice recognition system discarding the voice content;

step S07;

the voice recognition system selecting a called application program and writing a second function comprising a name and variables corresponding to the regular expressions with semantic variables into a private directory of the selected application program installed on the voice recognition system; and

step S08;

the system starting the selected application program, wherein the voice recognition system determines whether the private directory of the selected application program has the name of the second function or not before loading the selected application program, and wherein if the private directory of the selected application program has the name and variables of the second function, the selected application program loads the name and variables of the second function from the private directory and calls the second function; and

wherein if the private directory of the selected application program has no the name and the variables of the second function, a default initial page is loaded.

2. The construction method of claim 1, wherein the submitted information in step S01 comprises names and identifiers of the application programs, class information of the application programs, and a plurality of specific regular expressions with semantic variables and corresponding function call sets of the application program nodes and some information from the application program internal nodes.

3. The construction method of claim 1, wherein in step S07, a priority of each application program node that belongs to one of the class nodes is set based on the user'"'"'s settings, wherein if the regular expressions matching with the voice content does not designate that one of the application program nodes call the second function, a higher priority application program node is selected to call the second function.

4. The construction method of claim 1, wherein in step S07, only the selected application program and the voice recognition system can write data into the private directory of the selected application program installed on the voice recognition system, while only the selected application program can read the name and the variables of the second function that are stored in the private directory.

5. A voice recognition system, comprising:

a voice recognition module for converting voice content input by a user into text form;

a plurality of application programs for executing voice commands; and

a registration centre for storing information of the application programs, the registration centre comprising;

a plurality of class nodes each for storing a plurality of regular expressions with semantic variables which are used to describe common behaviors of a series of the application programs and mapping the regular expressions to a public function set;

wherein each of the application programs belongs to one of the class nodes and inherits the public function set of the class node, and is added into the voice recognition system via submitting a name, class information, and an identifier of each of the application programs to the registration centre, and wherein when text content output by the voice recognition module matches with the regular expressions with semantic variables of one class node, the voice recognition system selects and starts one of the application programs to execute an inherited public function set;

wherein the registration centre establishes a multi fork tree structure of the voice recognition system, and the voice recognition system further comprising;

a root node for storing a plurality of regular expressions with semantic variables to find and open files or the application programs;

a plurality of application program nodes for representing the application programs, and each application program node comprises a plurality of regular expressions with semantic variables that are not used in the class nodes and mapped to a specific function call set; and

a plurality of application program internal nodes for storing a plurality of regular expressions used in the application programs, and corresponding function sets.