×

SYSTEM AND METHOD FOR RAPID PROTOTYPING OF EXISTING SPEECH RECOGNITION SOLUTIONS IN DIFFERENT LANGUAGES

  • US 20100299133A1
  • Filed: 05/18/2010
  • Published: 11/25/2010
  • Est. Priority Date: 05/19/2009
  • Status: Active Grant
First Claim
Patent Images

1. A system for porting a speech recognition solution in a source language to recognize a target language, said speech recognition solution consisting of a speech recognition engine, pronunciation lexicon in the source language, speech grammar file for the source language, prompts in the source language, said system comprising:

  • lexicon conversion means adapted to convert the pronunciation lexicon of the source language to an equivalent lexicon to be used in the target language, said lexicon conversion means having;

    i. a first database for storing lexicon of words in the source language corresponding to analogous words in the target language;

    ii. translation means co-operating with said first database adapted to receive each word in the source language and provide its corresponding word in the target language;

    iii. transliteration means adapted to receive and map said translated word into source language graphemes;

    iv. grapheme to phoneme conversion means adapted to generate a source language phoneme sequence for each of said source language graphemes to obtain the phonetic pronunciation of the target language word in the source language;

    v. lookup table creation means adapted to receive a transliterated target language word in the source language and also receive equivalent phonetic pronunciation of said word from said grapheme to phoneme conversion means and prepare a lookup table;

    grammar conversion means adapted to modify the speech grammar file of the source language to handle free speech based speech recognition solutions in the target language, said grammar conversion means having;

    i. translation means adapted to receive the speech grammar file of the source language and translate the grammar file to the target language;

    ii. transliteration means adapted to receive said translated grammar file and transliterate said translated grammar file to the source language and provide a transliterated grammar file for the target language in the source language;

    prompt generation means adapted to convert voice prompts in the source language to the target language, said prompt generation means having;

    i. translation means adapted to convert words containable in the voice prompts of the source language into words in the target language;

    ii. identification means adapted to use the words in said lookup table of said lexicon conversion means to identify the phonetic pronunciation corresponding to the words containable in the prompts in the target language and provide a phoneme sequence of words containable in the prompts in the source language;

    iii. text to speech conversion means adapted to receive said phoneme sequence of words and generate a string of words containable in the prompts in the source language;

    iv. grammar adjustment means adapted to receive the converted string of words containable in the prompts in the source language and arrange the words in accordance with said transliterated grammar file for the target language and provide a grammatically modified text based prompt;

    v. speech generation means adapted to receive said grammatically modified text based prompt and generate speech based output for said grammatically modified text based prompt;

    prompting means co-operating with said prompt generation means adapted to output said generated speech corresponding to a prompt in the target language to guide users to submit their query;

    receiving means adapted to receive a speech based query from users in the target language, said query adapted to be received and converted into source language text and source language representation using said lookup table and said transliterated grammar file for the target language by the speech recognition engine;

    processing means adapted to process said source language text and perform pre-determined operation and further adapted to provide a processed output;

    compiling means co-operating with said lexicon conversion means, said speech grammar conversion means, said prompt generation means and said processing means and adapted to compile the final output in the target language; and

    playback means adapted to play said final output in the target language.

View all claims
  • 2 Assignments
Timeline View
Assignment View
    ×
    ×