Speech recognition using channel verification

• US 7,877,255 B2
• Filed: 03/31/2006
• Issued: 01/25/2011
• Est. Priority Date: 03/31/2006
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Voice Signal Technologies Incorporated (Microsoft Corporation)
• Inventors: Zlokarnik, Igor
• Primary Examiner(s): Smits; Talivaldis I
• Assistant Examiner(s): BORSETTI, GREG
• Application Number:
  US11/394,643
• Publication Number:
  US 20070239448A1
• Time in Patent Office: 1,761 Days
• Field of Search: 704/252, 704/E15.039, 704/256, 704/E15.05, 381/94.7, 381/94.1
• US Class Current: 704/252
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence

Abstract

A method for automatic speech recognition includes determining for an input signal a plurality scores representative of certainties that the input signal is associated with corresponding states of a speech recognition model, using the speech recognition model and the determined scores to compute an average signal, computing a difference value representative of a difference between the input signal and the average signal, and processing the input signal in accordance with the difference value.

Claims

1. A method for automatic speech recognition, the method comprising:

determining for an input signal a plurality of scores representative of certainties that the input signal is associated with corresponding states of a speech recognition model;

using the speech recognition model and the determined scores to compute an average signal;

computing, via a processor device executing instructions, a difference value representative of a difference between the input signal and the average signal; and

processing, via the processor device, the input signal in accordance with the difference value;

wherein computing the average signal comprises;

identifying a given score from the plurality of scores;

selecting from the plurality of scores a set of scores whose corresponding values are within a predetermined threshold from a value of the given score; and

performing an averaging operation on observation mean vectors of observation densities associated with the selected plurality of scores to obtain the average signal.

2. The method of claim 1, wherein processing the input signal in accordance with the difference value comprises:

performing speech recognition using the input signal if the difference value does not exceed a first predetermined value.

3. The method of claim 1, wherein processing the input signal in accordance with the difference value comprises:

adjusting the plurality of scores based on the difference value; and

using the adjusted scores to update the speech recognition model.

4. The method of claim 3, further comprising:

applying to the input signal a cepstrum transformation; and

generating a mel-frequency cepstral representation comprising mel-frequency coefficients from the cepstrum transform representation of the input signal.

5. The method of claim 4, wherein the average signal is represented using a corresponding mel-frequency cepstral representation, and wherein computing the difference value comprises:

computing a channel deviation value based on a difference between a 0^th dimension of a mel-frequency cepstral coefficient of the mel-cepstral frequency representation of input signal and a 0^th dimension mel-frequency cepstral representation of the average signal.

6. The method of claim 1, wherein determining the plurality of scores includes:

performing a negative logarithmic operation on each of the certainties associated with the corresponding states.

14. The method as in claim 1 further comprising:

segmenting the input signal into frames;

computing a respective coefficient for each of the frames;

normalizing the coefficients associated with the frames;

presenting the normalized coefficients to a speech recognition decoder; and

computing the plurality of scores using the normalized coefficients.

15. The method as in claim 14, wherein the scores are probability scores, the method further comprising:

selecting a subset of scores from the probability scores;

generating the average signal based on the subset of scores; and

adjusting the plurality of scores based on the average signal.

16. The method as in claim 15 further comprising:

utilizing the adjusted plurality of scores to determine whether the input signal represents sound on a noise channel or sound on a speech channel.

17. The method as in claim 1, wherein processing the input signal in accordance with the difference value includes:

adjusting the plurality of scores based on the average signal; and

utilizing the adjusted plurality of scores to determine whether or not the input signal represents sound on a noise channel or sound on a speech channel.

18. The method as in claim 1 further comprising:

determining that the input signal represents sound on a noise channel, as opposed to sound on a speech channel, based on a magnitude of the difference value.

19. The method as in claim 1 further comprising:

in response to determining that a segment of the input signal is inconsistent with signals normally received through a speech channel, biasing scores of noise states associated with the speech recognition model to increase a probability that the segment of the input signal is deemed to be sound received on a noise channel rather than sound received on the speech channel.

20. The method as in claim 19 further comprising:

detecting that the segment is received on the noise channel in response to detecting that the difference value is above a threshold.

21. The method as in claim 1, wherein the scores are probability scores associated with states of the input signal and wherein the average signal is a moving average generated based on the input signal, the method further comprising:

biasing scores of noise states associated with the speech recognition model to increase a probability that a segment of the input signal is deemed to be sound received on a noise channel rather than sound received on a speech channel.

7. A method for automatic speech recognition, the method comprising:

determining for an input signal a plurality of scores representative of certainties that the input signal is associated with corresponding states of a speech recognition model;

using the speech recognition model and the determined scores to compute an average signal;

computing, via a processor device executing instructions, a difference value representative of a difference between the input signal and the average signal;

processing, via the processor device, the input signal in accordance with the difference value;

wherein processing the input signal in accordance with the difference value comprises;

performing speech recognition using the input signal if the difference value does not exceed a first predetermined threshold; and

wherein computing the average signal comprises;

identifying from the plurality of scores a best score;

selecting from the plurality of scores those scores whose corresponding values are within a second predetermined threshold from a value of the best score; and

performing an averaging operation on observation mean vectors of observation densities associated with the selected scores to obtain the average signal.

8. A computer program product for automatic speech recognition, the computer program product stored on a non-transitory machine readable storage medium for storing computer instructions that, when executed, cause a processor-based machine to:

determine for an input signal a plurality of scores representative of certainties that the input signal is associated with corresponding states of a speech recognition model;

use the speech recognition model and the determined scores to compute an average signal;

compute a difference value representative of a difference between the input signal and the average signal; and

process the input signal in accordance with the difference value;

execution of the instructions further causing the processor-based machine to compute the average signal via steps of;

identifying a given score from the plurality of scores;

selecting from the plurality of scores a set of scores whose corresponding values are within a predetermined threshold from a value of the given score; and

performing an averaging operation on observation mean vectors of observation densities associated with the selected plurality of scores to produce the average signal.

9. The computer program product of claim 8, wherein the instructions that cause the processor-based machine to process the input signal in accordance with the difference value comprise instructions that, when executed, cause the processor-based machine to:

perform speech recognition using the input signal if the difference value does not exceed a first predetermined value.

10. The computer program product of claim 8, wherein the instructions that cause the processor-based machine to process the input signal in accordance with the difference value comprise instructions that, when executed, cause the processor-based machine to:

adjust the plurality of scores based on the difference value; and

use the adjusted scores to update the speech recognition model.

11. The computer program product of claim 10, further comprising instructions that when executed cause the processor-based machine to:

apply to the input signal a cepstrum transformation; and

generate a mel-frequency cepstral representation comprising mel-frequency coefficients from the cepstrum transform representation of the input signal.

12. The computer program product of claim 11, wherein the average signal is represented using a corresponding mel-frequency cepstral representation, and wherein the instructions that that cause the processor-based machine to compute the difference value comprise instructions that when executed cause the processor-based machine to:

compute a channel deviation value based on a difference between a 0^th dimension of a mel-frequency cepstral coefficient of the mel-cepstral frequency representation of input signal and a 0^th dimension mel-frequency cepstral representation of the average signal.

13. A computer program product for automatic speech recognition, the computer program product stored on a non-transitory machine readable storage medium for storing computer instructions that, when executed, cause a processor-based machine to:

determine for an input signal a plurality of scores representative of certainties that the input signal is associated with corresponding states of a speech recognition model;

use the speech recognition model and the determined scores to compute an average signal;

compute a difference value representative of a difference between the input signal and the average signal; and

process the input signal in accordance with the difference value;

wherein the instructions that cause the processor-based machine to compute the average signal comprise instructions that, when executed, cause the processor-based machine to;

identify from the plurality of scores a best score;

select from the plurality of scores those scores whose corresponding values are within a second predetermined threshold from a value of the best score; and

perform an averaging operation on observation mean vectors of observation densities associated with the selected scores to obtain the average signal.