Speech recognition device using statistical language model
The present invention includes specific sentence generator 14 for generating a piece of sentence data specific to an application from a grammar stored in grammar storing unit 12 and correction information stored in correction information storing unit 13 and then storing the piece of data in specific sentence storing unit 15 , topic shift unit 16 for shifting a general topic stored in general topic storing unit 17 to a topic specific to an application on the basis of the piece of sentence data and then storing the topic in shifted topic storing unit 18 , natural speech recognition unit 21 for recognizing speech by dictation with a shifted topic and then storing the recognition result in recognition result storing unit 22 , candidate sentence generator 19 for generating candidate sentences from a grammar and then storing the sentences in candidate sentence storing unit 20 , and matching unit 23 for determining a final recognition result by matching candidate sentences with a recognition result.
1. A speech recognizing device comprising:
natural speech recognizing means for recognizing speech input in an application program by dictation; and
recognition result converting means for converting a recognition result from said natural speech recognizing means into a final recognition result processable by said application program on the basis of a grammar to be used for recognizing said input speech in a grammar method, wherein said recognition result converting means further comprises:
candidate sentence generating means for evolving said grammar to generate candidate sentences that are candidates for said final recognition result; and
matching means for selecting a candidate sentence as said final recognition result among the candidate sentences by matching said candidate sentences generated by said candidate sentence generating means against the recognition result by said natural speech recognizing means, and wherein the candidate sentence comprises a sentence having an associated highest score calculated according to the formula:
(ngramWords—replacedWords—rejectedWords—insertedWords)/ngramWords
wherein “ngramWords” comprises a first number of words contained in an N-gram recognition result, “replacedWords” comprises a second number of words that have been replaced with different words for the N-gram recognition result, “rejectedWords” comprises a third number of words that are contained in the N-gram recognition result but not in the candidate sentences, and “insertedWords” comprises a fourth number of words that are not contained in the N-gram recognition result but are contained in the candidate sentences.
2. The speech recognizing device according to claim 1 , further comprising statistical language model generating means for generating a statistical language model to be used for recognizing said speech inputted to an application program by dictation on the basis of a grammar to be used for recognizing speech inputted to the application program in grammar method; and
wherein said natural speech recognizing means recognizes said inputted speech by dictation using said statistical language model generated by said statistical language model generating means.
3. The speech recognizing device according to claim 2 , wherein said statistical language model generating means modifies a general statistical language model based on said grammar in order to generate said statistical language model.
4. The speech recognizing device according to claim 2 , wherein said statistical language model generating means generates said statistical language model separately from a general statistical language model; and
said natural speech recognizing means recognizes said inputted speech by dictation using said general statistical language model and said statistical language model generated by said statistical language model generating means.
5. A computer program product on a computer readable medium, the computer program product for causing a computer to implement the operations of:
receiving recognition speech input in an application program by dictation; and
converting a recognition result by said dictation into a final recognition result processable by said application program on the basis of a grammar to be used for recognizing said input speech in a grammar method, wherein converting further comprises:
evolving said read-out grammar to generate candidate sentences that are candidates for said final recognition result; and
selecting a candidate sentence as said final recognition result among the candidate sentences by matching said generated candidate sentences against the recognition result by said dictation, and wherein the candidate sentence comprises a sentence having an associated highest score calculated according to the formula:
(ngramWords—replacedWords—rejectedWords—insertedWords)/ngramWords
wherein “ngramWords” comprises a first number of words contained in an N-gram recognition result, “replacedWords” comprises a second number of words that have been replaced with different words for the N-gram recognition result, “rejectedWords” comprises a third number of words that are contained in the N-gram recognition result but not in the candidate sentences, and “insertedWords”comprises a fourth number of words that are not contained in the N-gram recognition result but are contained in the candidate sentences.