Non-interactive enrollment in speech recognition
First Claim
1. A computer-implemented method for enrolling a user in a speech recognition system, comprising:
- obtaining data representing a user'"'"'s speech, the speech including multiple user utterances and generally corresponding to an enrollment text;
analyzing acoustic content of data corresponding to a user utterance;
determining, based on the analysis, whether the user utterance matches a portion of the enrollment text; and
if the user utterance matches a portion of the enrollment text, using the acoustic content of the user utterance to update acoustic models corresponding to the portion of the enrollment text,wherein determining whether the user utterance matches a portion of the enrollment text comprises a determination that the user utterance matches when the user has skipped at least one word of the portion of the enrollment text.
8 Assignments
0 Petitions
Accused Products
Abstract
A computer enrolls a user in a speech recognition system by obtaining data representing a user'"'"'s speech, the speech including multiple user utterances and generally corresponding to an enrollment text, and analyzing acoustic content of data corresponding to a user utterance. The computer determines, based on the analysis, whether the user utterance matches a portion of the enrollment text. If so, the computer uses the acoustic content of the user utterance to update acoustic models corresponding to the portion of the enrollment text. The computer may determine that the user utterance matches a portion of the enrollment text even when the user has skipped or repeated words of the enrollment text.
210 Citations
31 Claims
-
1. A computer-implemented method for enrolling a user in a speech recognition system, comprising:
-
obtaining data representing a user'"'"'s speech, the speech including multiple user utterances and generally corresponding to an enrollment text; analyzing acoustic content of data corresponding to a user utterance; determining, based on the analysis, whether the user utterance matches a portion of the enrollment text; and if the user utterance matches a portion of the enrollment text, using the acoustic content of the user utterance to update acoustic models corresponding to the portion of the enrollment text, wherein determining whether the user utterance matches a portion of the enrollment text comprises a determination that the user utterance matches when the user has skipped at least one word of the portion of the enrollment text. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. Computer software, residing on a computer-readable storage medium, comprising instructions for causing a computer to:
-
obtain data representing a user'"'"'s speech, the speech including multiple user utterances and generally corresponding to an enrollment text; analyze acoustic content of data corresponding to a user utterance; determine, based on the analysis, whether the user utterance matches a portion of the enrollment text; and
use the acoustic content of the user utterance to update acoustic models corresponding to a portion of the enrollment text that matches the user utterance;wherein the instructions configure the computer to determine that the user utterance matches when the user has skipped at least one word of the portion of the enrollment text. - View Dependent Claims (22, 23)
-
-
24. A speech recognition system for enrolling a user, comprising:
-
an input device for receiving speech signals; and
a processor configured to;
obtain data representing a user'"'"'s speech, the speech including multiple user utterances and generally corresponding to an enrollment text;analyze acoustic content of data corresponding to a user utterance; determine, based on the analysis, whether the user utterance matches a portion of the enrollment text;
use the acoustic content of the user utterance to update acoustic models corresponding to a portion of the enrollment text that matches the user utterance; anddetermine that the user utterance matches when the user has skipped at least one word of the portion of the enrollment text. - View Dependent Claims (25, 26)
-
-
27. A computer-implemented method for enrolling a user in a speech recognition system, comprising:
-
obtaining data representing a user'"'"'s speech from a recording device physically separate from the computer implementing the method, the speech including multiple user utterances and generally corresponding to an enrollment text; analyzing acoustic content of the obtained data corresponding to a user utterance to identify a sequence of words in the user utterance; determining, using the sequence of words, whether the user utterance matches a portion of the enrollment text; and if the user utterance matches a portion of the enrollment text, using the acoustic content of the user utterance to update acoustic models corresponding to the portion of the enrollment text, wherein determining whether the user utterance matches a portion of the enrollment text comprises a determination that the user utterance matches when the sequence of words in the user utterance is different than a sequence of words in the portion of the enrollment text. - View Dependent Claims (28, 29, 30, 31)
-
Specification