Back to results
US 11,145,292 B2Granted
Method and device for updating language model and performing speech recognition based on language model
AssigneeSamsung Electronics Co., Ltd.
InventorsDong-hoon Ahn, Chi-youn Park, Kyung-min Lee, Jae-won Lee
Life of this patent
Estimating the term…
Abstract
A method of updating a grammar model used during speech recognition includes obtaining a corpus including at least one word, obtaining the at least one word from the corpus, splitting the at least one obtained word into at least one segment, generating a hint for recombining the at least one segment into the at least one word, and updating the grammar model by using at least one segment comprising the hint.
Claims9
A method of recognizing a voice signal, the method comprising:
- obtaining a first voice signal and a second voice signal;
- recognizing the first voice signal by applying a first voice recognition model to the obtained first voice signal;
- based on at least the applying of the first voice recognition model, determining whether the first voice signal includes a signal corresponding to a word from among a plurality of predefined words;
- selecting a second voice recognition model from among a plurality of second voice recognition models, based on at least the determining that the first voice signal includes the signal corresponding to the word, each of the plurality of the second voice recognition models corresponding to at least one respective predefined word of the plurality of predefined words;
- recognizing the second voice signal based on the selected second voice recognition model;
- determining whether the second voice signal includes a signal corresponding to another word from among the plurality of predefined words or a signal corresponding to an exit token associated with the selected second voice recognition model, based on the recognition result for the second voice signal;
- selecting the first voice recognition model, based on at least the determining that the second voice signal includes the signal corresponding to the exit token associated with the selected second voice recognition model;
- selecting a third voice recognition model from among the plurality of second voice recognition models, based on at least the determining that the second voice signal includes the signal corresponding to the other word;
- and
- recognizing a third voice signal, which is obtained after the second voice signal, based on the selected first voice recognition model or the selected third voice recognition model.
7 more elements
A device for recognizing a voice signal, the device comprising:
- a receiver for obtaining a first voice signal and a second voice signal;
- a memory for storing a first voice recognition model and a plurality of second voice recognition models;
- and
- at least one processor configured for:
- recognizing the first voice signal by applying the first voice recognition model to the obtained first voice signal,
- determining, based on the applying of the first voice recognition model, whether the first voice signal includes a signal corresponding to a word from among a plurality of predefined words,
- selecting a second voice recognition model from among the plurality of second voice recognition models, based on the determining that the first voice signal includes the signal corresponding to the word, each of the plurality of the second voice recognition models corresponding to at least one respective predefined word of the plurality of predefined words, and
- recognizing the second voice signal based on the selected second voice recognition model,
- determining whether the second voice signal includes a signal corresponding to another word from among the plurality of predefined words or a signal corresponding to an exit token associated with the selected second voice recognition model, based on the recognition result for the second voice signal,
- selecting the first voice recognition model, based on at least the determining that the second voice signal includes the signal corresponding to the exit token associated with the selected second voice recognition model,
- selecting a third voice recognition model from among the plurality of second voice recognition models, based on at least the determining that the second voice signal includes the signal corresponding to the other word, and
- recognizing a third voice signal, which is obtained after the second voice signal, based on the selected first voice recognition model or the selected third voice recognition model.
9 more elements
Dependencies are inferred from the claim text.