Distinguishing out-of-vocabulary speech from in-vocabulary speech
A speech recognition method includes receiving input speech from a user, processing the input speech using a first grammar to obtain parameter values of a first N-best list of vocabulary, comparing a parameter value of a top result of the first N-best list to a threshold value, and if the compared parameter value is below the threshold value, then additionally processing the input speech using a second grammar to obtain parameter values of a second N-best list of vocabulary. Other preferred steps include: determining the input speech to be in-vocabulary if any of the results of the first N-best list is also present within the second N-best list, but out-of-vocabulary if none of the results of the first N-best list is within the second N-best list; and providing audible feedback to the user if the input speech is determined to be out-of-vocabulary.
1. A speech recognition method comprising the steps of:
(a) receiving input speech from a user via a microphone associated with an automatic speech recognition system;
(b) processing the input speech using a first grammar to obtain parameter values of a first N-best list of vocabulary using at least one processor associated with the automatic speech recognition system;
(c) comparing at least one parameter value of a top result of the first N-best list to at least one predetermined threshold value; and
(d) subsequently processing the input speech using a second grammar to obtain parameter values of a second N-best list of vocabulary, if the compared at least one parameter value is below the at least one predetermined threshold value;
(e) determining the input speech to be in-vocabulary, if any of the results of the first N-best list is also present within the second N-best list, but out-of vocabulary if any of the results of the first N-best list is not within the second N-best list; and
(f) providing audible feedback to the user if the input speech is determined to be out-of-vocabulary.
2. The method of claim 1 , wherein the at least one parameter value is at least one of a confidence score, segment duration, likelihood score, or signal-to-noise ratio.
3. The method of claim 1 , wherein step (d) is disabled if frequency of out-of-vocabulary utterances falls below a predetermined threshold.
4. The method of claim 3 , wherein the top result from step (c) is determined to be the vocabulary contained in the user's speech.
5. The method of claim 1 , wherein step (d) is enabled if the user is a novice to the system.
6. The method of claim 1 , wherein step (d) is enabled for predetermined types of grammar.
7. The method of claim 1 , further comprising the step of creating the second grammar by generating a limited set of vocabulary that includes in-vocabulary within the first grammar.
8. The method of claim 1 , further comprising the step of creating the second grammar by combining two or more other grammars used by a speech recognition system used to implement the method.
9. The method of claim 1 , further comprising the step of creating the second grammar by generating a limited set of vocabulary that does not include in-vocabulary of the first grammar.
10. The method of claim 1 , further comprising the step of creating the second grammar by generating a limited set of vocabulary that does not include triphones or pentaphones of words of the in-vocabulary first grammar.
11. The method of claim 7 , wherein step (e) includes determining the input speech to be in-vocabulary, if one of the top results of the first N-best list is also present within the second N-best list, but out-of-vocabulary if none of the top results of the first N-best list is within the second N-best list.
12. The method of claim 11 , wherein step (e) includes determining the input speech to be in-vocabulary, if the top result of the first N-best list is also present within the second N-best list, but out-of-vocabulary if the top result of the first N-best list is not within the second N-best list.
13. The method of claim 11 , wherein step (e) includes determining the input speech to be in-vocabulary, if the top result of the first N-best list is also present within the top results of the second N-best list, but out-of-vocabulary if the top result of the first N-best list is not within the top results of the second N-best list.
14. The method of claim 1 , wherein the step (f) includes instructing the user to use vocabulary within the first grammar.
15. The method of claim 14 , wherein the step (f) further includes providing a list of acceptable vocabulary within the first grammar.
16. The method of claim 1 , further comprising determining whether latency between steps (b) and (d) falls outside a defined period and, if so, skipping step (d) and redefining the first grammar to also include the second grammar.
17. A speech recognition method for a telematics system that has a speech recognition system, the method comprising the steps of:
(a) receiving input speech from a user to be processed into recognized speech via a microphone associated with an automatic speech recognition system;
(b) processing the input speech using a first grammar to obtain parameter values of a first N-best list of probable vocabulary suggested by a decoder using at least one processor associated with the automatic speech recognition system;
(c) comparing at least one parameter value of a top result of the first N-best list to at least one predetermined threshold value;
(d) subsequently processing the input speech using a second grammar to obtain parameter values of a second N-best list of probable vocabulary suggested by the decoder, if the compared at least one parameter value is below the at least one predetermined threshold value; and
(e) determining the input speech to be in-vocabulary, if any of the results of the first N-best list is also present within the second N-best list, but out-of-vocabulary if none of the results of the first N-best list is within the second N-best list,
wherein the first grammar is different from the second grammar, and the first grammar includes in-vocabulary grammar and the second grammar includes out-of-vocabulary grammar; and
(f) providing audible feedback to the user if the input speech is determined to be out-of-vocabulary.
18. A speech recognition method for a vehicle having a telematics unit with an embedded speech recognition system, the method comprising the steps of:
(a) receiving input speech from a user to be processed into recognized speech via a microphone associated with an automatic speech recognition system;
(b) processing the input speech using a first grammar to obtain parameter values of a first N-best list of probable vocabulary suggested by a decoder using at least one processor associated with the automatic speech recognition system;
(c) comparing at least one parameter value of a top result of the first N-best list to at least one predetermined threshold value, wherein the at least one parameter value is at least one of a confidence score, segment duration, likelihood score, or signal-to-noise ratio;
(d) subsequently processing the input speech using a second grammar to obtain parameter values of a second N-best list of probable vocabulary suggested by the decoder, if the compared at least one parameter value is below the at least one predetermined threshold value;
(e) determining the input speech to be in-vocabulary, if the top result of the first N-best list is also present within the second N-best list, but out-of-vocabulary if the top result of the first N-best list is not within the second N-best list; and
(f) providing audible feedback to the user if the input speech is determined to be out-of-vocabulary,
wherein the first grammar is different from the second grammar, and the first grammar includes in-vocabulary grammar and the second grammar includes out-of-vocabulary grammar.