Method of recognizing speech

1. A method of recognizing speech, comprising:

presenting an utterance to a speech recognition system;

determining, via the speech recognition system, that the utterance contains a particular expression, the particular expression capable of being associated with at least two different meanings;

splitting the utterance into a plurality of speech frames, each frame being assigned a predetermined time segment and a frame number;

indexing the utterance to i) a predetermined frame number, or ii) a predetermined time segment, the indexing identifying that one of the plurality of frames includes the particular expression; and

re-presenting the one of the plurality of frames including the particular expression to the speech recognition system to verify that the particular expression was actually recited in the utterance.

2. The method as defined in claim 1 wherein prior to the splitting, the method further comprises:

generating an N-best list for the recited utterance, the N-best list including a plurality of possible utterances, each of which potentially corresponds with the recited utterance, wherein each of the plurality of possible utterances is ranked according to a confidence score; and

identifying a first-best possible utterance from the N-best list, the first-best possible utterance being ranked the highest among the plurality of possible utterances according to its confidence score, wherein the first-best possible utterance contains one meaning of the particular expression.

3. The method as defined in claim 2 wherein the one of the plurality of frames includes a digitized form of the particular expression.

4. The method as defined in claim 2 wherein upon verifying that the particular expression was actually recited in the utterance, the method further comprises:

identifying an other possible utterance from the N-best list, the other possible utterance including an other expression appended to the particular expression that associates an other meaning with the particular expression;

presenting at least one frame preceding the one of the plurality of frames including the particular expression to the speech recognition system to verify that the other expression was actually recited in the utterance;

swapping the first-best possible utterance with the other possible utterance from the N-best list when the other expression is verified; and

presenting the other possible utterance to a source of the utterance.

5. The method as defined in claim 1 wherein prior to splitting the plurality of speech frames, the method further comprises:

generating an N-best list for the recited utterance, the N-best list including a plurality of possible utterances, each of which potentially corresponds with the recited utterance, wherein each of the plurality of possible utterances is ranked according to a confidence score;

identifying a first-best possible utterance from the N-best list, wherein the first-best possible utterance does not contain the particular expression; and

presenting the first-best possible utterance to a source of the utterance.

6. The method as defined in claim 1 wherein after verifying that the particular expression was actually recited in the utterance, the method further comprises determining whether speech energy is present in the utterance before the one of the plurality of frames including the particular expression, the presence of the speech energy indicating that one meaning of the particular expression is present in the utterance.

7. The method as defined in claim 6 , further comprising prompting a source of the utterance to verify that the one meaning of the particular expression is accurate.

8. The method as defined in claim 1 wherein the utterance is derived from a tonal language, and wherein an English equivalent of the two different meanings of the particular expression include “yes” and “no”.

9. The method as defined in claim 2 wherein upon verifying that the particular expression was actually recited in the utterance, the method further comprises:

recognizing that an other possible utterance from the N-best list does not contain an other meaning of the particular expression; and

presenting the first-best possible utterance from the N-best list to a source of the utterance.

10. A speech for recognizing speech, comprising:

a microphone configured to receive an utterance from a source, the utterance potentially containing a particular expression, wherein the particular expression is capable of being associated with at least two different meanings;

a processor selectively and operatively connected to the microphone, the processor including:

computer readable code for determining that the utterance contains the particular expression;

computer readable code for splitting the utterance into a plurality of speech frames, each frame being assigned a predetermined time segment and a frame number; and

computer readable code for indexing the utterance to i) a predetermined frame number, or ii) a predetermined time segment, the indexing identifying that one of the plurality of frames includes the particular expression; and

means for re-presenting the one of the plurality of frames including the particular expression to the speech recognition system to verify that the particular expression was actually recited in the utterance.

11. The system as defined in claim 10 wherein the processor further includes:

computer readable code for generating an N-best list for the utterance, the N-best list including a plurality of possible utterances, each of which potentially corresponds with the utterance, wherein each of the plurality of possible utterances is ranked according to a confidence score; and

computer readable code for identifying a first-best possible utterance from the N-best list, the first-best possible utterance being ranked the highest among the plurality of possible utterances according to its confidence score, wherein the first-best possible utterance contains one meaning of the particular expression.

12. The system as defined in claim 11 wherein the processor further includes:

computer readable code for identifying an other possible utterance from the N-best list, the other possible utterance including an other expression appended to the particular expression that associates an other meaning with the particular expression;

computer readable code for presenting at least one frame preceding the one of the plurality of frames including the particular expression to the speech recognition system to verify that the other expression was actually recited in the utterance; and

computer readable code for swapping the first-best possible utterance with the other possible utterance from the N-best list when the other expression is verified.

13. The system as defined in claim 12 , further comprising an audio component configured to present, in response to a command from the processor, the other possible utterance to a source of the utterance.

14. The system as defined in claim 11 wherein the processor further includes computer readable code for recognizing that an other possible utterance from the N-best list does not contain an other meaning of the particular expression.

15. The system as defined in claim 10 wherein the processor further includes:

computer readable code for generating an N-best list for the utterance, the N-best list including a plurality of possible utterances, each of which potentially corresponds with the utterance, wherein each of the plurality of possible utterances is ranked according to a confidence score; and

computer readable code for identifying a first-best possible utterance from the N-best list, wherein the first-best possible utterance does not contain the particular expression.

16. The system as defined in claim 15 , further comprising an audio component configured to present, in response to a command from the processor, the first-best possible utterance to a source of the utterance.

17. The system as defined in claim 10 wherein the processor further includes computer readable code for determining whether speech energy is present in the utterance before the one of the plurality of frames including the particular expression, the presence of the speech energy indicating that one meaning of the particular expression is present in the utterance.

18. The system as defined in claim 17 , further comprising an audio component configured to prompt a source of the utterance to verify that the one meaning of the particular expression is accurate.

19. The system as defined in claim 10 wherein the utterance is derived from a tonal language, and wherein an English equivalent of the two different meanings of the particular expression includes “yes” and “no”.

20. The system as defined in claim 10 wherein the system is configured to be used in a mobile vehicle.