A kind of speech recognition system and construction method

A kind of speech recognition system and construction method

  • CN 103,400,579 B
  • Filed: 08/04/2013
  • Issued: 11/18/2015
  • Est. Priority Date: 08/04/2013
  • Status: Active Grant
First Claim
Patent Images

1. a speech recognition system construction method, described speech recognition system comprises sound identification module and registration center, it is characterized in that:

  • comprise the steps,S01;

    application program submits to information to registration center;

    S02;

    the phonetic entry of system receives user, sound identification module is converted into textual form voice content;

    S03;

    the voice content of system verification input whether with the matching regular expressions of the band semantic variant of root node, if coupling just performs corresponding function call, otherwise perform next step;

    S04;

    voice content is delivered to category node, if the regular expression of the band semantic variant of voice content and all category nodes does not mate, then performs step S05, otherwise performs step S07;

    S05;

    voice content is delivered to all application program nodes and carries out semantics recognition, if the voice content of input and the matching regular expressions of described application program node, system just performs step S07, if the voice content of input and the regular expression of all application program nodes do not mate just perform step S06;

    S06;

    system abandons voice content or voice content input public search engines is searched on a large scale;

    S07;

    Systematic selection by invoked application program, and is in the privately owned catalogue of selected application program foundation the function call writing system of the regular expression corresponding to band semantic variant;

    S08;

    system starts selected application program, application program can judge whether have function and corresponding variable in the privately owned catalogue of application program before loading, if exist, application program will from privately owned catalogue loading function and variable, and carry out calling of respective function, otherwise load default initial page;

    Wherein, described registration center, install application information in systems in which in order to storage and set up the speech recognition system structural drawing of a Multiway Tree Structure, it comprises;

    Described root node, stores the regular expression of band semantic variant, performs and searches and open storage file in systems in which or the operation of application program;

    Described category node, stores the regular expression having the band semantic variant of behavior for describing a series of application program, and these regular expressions are mapped in an open collection of functions;

    Described application program node, belong to certain category node and inherit the application program of the open collection of functions of affiliated category node, also there is the regular expression of some category nodes unexistent band semantic variant and be mapped to specific function call in described application program node simultaneously;

    Application program internal node, stores the regular expression and corresponding function call that are applied to application program inside.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×