Tailored speaker-independent voice recognition system
View Patent ↗A tailored speaker-independent voice recognition system has a speech recognition dictionary ( 360 ) with at least one word ( 371 ). That word ( 371 ) has at least two transcriptions ( 373 ), each transcription ( 373 ) having a probability factor ( 375 ) and an indicator ( 377 ) of whether the transcription is active. When a speech utterance is received ( 510 ), the voice recognition system determines ( 520, 530 ) the word signified by the speech utterance, evaluates ( 540 ) the speech utterance against the transcriptions of the correct word, updates ( 550 ) the probability factors for each transcription, and inactivates ( 570 ) any transcription that has an updated probability factor that is less than a threshold.
1. An electronic device comprising:
a memory for storing a word, a first transcription of the word, and a second transcription of the word;
a microphone for capturing a speech utterance;
a speech recognition engine, coupled to the microphone and the memory, for evaluating through a predetermined number of times of speech utterances against the first transcription and the second transcription, for determining a first probability factor for the first transcription based on an accumulated acoustic score for the first transcription and a total accumulated score of the first transcription and the second transcription, and for determining a second probability factor for the second transcription based on an accumulated acoustic score for the second transcription and the total accumulated score of the first transcription and the second transcription; and
a processor, coupled to the speech recognition engine, for inactivating the first transcription if the first probability factor is below a threshold.
2. An electronic device according to claim 1 further comprising:
a transcription generator, coupled to the processor, for automatically generating the first transcription.
3. An electronic device according to claim 2 wherein the transcription generator automatically generates the second transcription.
4. An electronic device according to claim 1 further comprising:
a display, coupled to the processor, for displaying the word.
5. An electronic device according to claim 1 further comprising:
a loudspeaker, coupled to the processor, for announcing the word.
6. An electronic device according to claim 1 wherein the memory further comprises:
a speech recognition dictionary for storing the word, the first transcription, the second transcription, the first probability factor, and the second probability factor.
7. An electronic device according to claim 6 wherein the memory further comprises:
a first transcription “active” indicator and a second transcription“active” indicator.
8. An electronic device according to claim 1 wherein the memory further comprises:
an electronic phonebook.
9. An electronic device according to claim 1 wherein the threshold is a predetermined value.
10. An electronic device according to claim 1 wherein the threshold is a dynamically-determined value.
11. An electronic device according to claim 10 wherein the threshold is a value equivalent to an x-th highest probability factor for the word.
12. An electronic device according to claim 10 wherein the threshold is a value equivalent to a predetermined fraction of a highest probability factor for the word.
13. A method for recognizing a speech utterance comprising the steps of:
obtaining a word;
receiving a first transcription for the word;
receiving a second transcription for the word;
initializing a first probability factor for the first transcription;
initializing a second probability factor for the second transcription;
activating the first transcription;
activating the second transcription;
receiving speech utterances;
determining the word signified by the speech utterances;
evaluating through a predetermined number of times the speech utterances against the first transcription;
evaluating through a predetermined number of times the speech utterances against the second transcription;
updating the first probability factor based on an accumulated acoustic score for the first transcription and a total accumulated score of the first transcription and the second transcription;
updating the second probability factor based on an accumulated acoustic score for the second transcription and the total accumulated score of the first transcription and the second transcription; and
inactivating the first transcription if the updated first probability factor is less than a threshold.
14. A method according to claim 13 wherein the threshold is a predetermined value.
15. A method according to claim 13 wherein the threshold is a dynamically determined value.
16. A method according to claim 13 wherein the step of receiving the first transcription comprises:
automatically creating the first transcription from the word.
17. A method according to claim 13 wherein the step of receiving the first transcription comprises:
storing the first transcription for the word in a memory.