Speech-recognition grammar analysis

• US 7,243,071 B1
• Filed: 01/16/2003
• Issued: 07/10/2007
• Est. Priority Date: 01/16/2003
• Status: Active Grant

• Current Assignee: Mavenir Systems Incorporated (Mavenir Group Holdings Ltd.)
• Original Assignee: Comverse Incorporated (Mavenir Group Holdings Ltd.)
• Inventors: Zinovieva, Nina, Resnick, David I.
• Primary Examiner(s): Hudspeth; David
• Assistant Examiner(s): Rider; Justin W
• Application Number:
  US10/345,853
• Time in Patent Office: 1,636 Days
• Field of Search: 704/231, 704/257
• US Class Current: 704/257
• 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

An apparatus for determining at least one recognition parameter, associated with a grammar, for use in attempting to recognize utterances using the grammar includes an input configured to receive indicia of the grammar, an output configured to provide the at least one recognition parameter, and a processor coupled to the input and to the output and configured to analyze words of the grammar to determine a feature of the grammar and to provide indicia of the at least one recognition parameter to the output based on the determined feature.

Claims

1. An apparatus for determining at least one recognition parameter, associated with a grammar, for use in attempting to recognize utterances using the grammar, the apparatus comprising:

an input receiving indicia of the grammar;

an output providing the at least one recognition parameter;

a processor coupled to the input and to the output;

a memory including computer executable instructions that, when executed by the processor, cause the processor to;

analyze words of the grammar to determine a feature of the grammar wherein the feature includes at least one of perplexity, word length, confusability, and prevalence of numbers associated with the grammar;

provide indicia of the at least one recognition parameter to the output based on the determined feature;

determine at least the perplexity and the word length associated with the grammar; and

determine a category of the grammar based on the determined perplexity and word length.

2. The apparatus of claim 1 wherein the computer readable instructions, when executed, further cause the processor to provide at least one predetermined recognition parameter value based on the determined category of the grammar.

3. The apparatus of claim 1 wherein the computer readable instructions, when executed, further cause the processor to determine confusability and prevalence of numbers.

4. A computer program product for use with a grammar to determine at least one recognition parameter of the grammar for use in attempting to recognize utterances using the grammar, the computer program product residing on a computer-readable medium and comprising computer-readable and computer-executable instructions for causing the computer to:

receive indicia of the grammar;

analyze words of the grammar to determine a feature of the grammar; and

output indicia of the at least one recognition parameter based on the determined feature;

wherein the feature includes at least one of perplexity, word length, confusability, and prevalence of numbers associated with the grammar;

wherein the instructions for causing the computer to analyze words cause the computer to determine at least the perplexity and the word length associated with the grammar; and

wherein the instructions for causing the computer to analyze words cause the computer to determine a category of the grammar based on the determined perplexity and word length.

5. The computer program product of claim 4 wherein the instructions for causing the computer to output indicia cause the computer to provide at least one predetermined recognition parameter value based on the determined category of the grammar.

6. An Interactive Voice Response (IVR) system comprising:

an input for receiving an utterance;

a first module to analyze portions of a grammar to determine at least one feature of the grammar and to provide at least one potential speech recognition parameter value in accordance with the at least one feature; and

a second module coupled to the first module and the input, the second module applies a speech model in accordance with the at least one potential speech recognition parameter value to attempt to recognize the utterance;

wherein the at least one feature includes at least one of perplexity, average word length, confusability, and relative prevalence of numbers in the grammar;

wherein the at least one feature includes at least perplexity and average word length;

wherein the at least one potential speech recognition parameter includes indicia of at least one of;

whether to use relatively high or low word-internal acoustic evaluation, at least one timeout length, a confidence level, a pruning amount, noise reduction, whether to use multiple acoustic models, and whether to use relatively high or low word-boundary acoustic evaluation; and

wherein the first module categorizes the grammar into one of four categories based on whether the perplexity is high or low and whether the average word length is long or short.

7. The system of claim 6 wherein the first module is configured to provide predetermined values of the at least one potential speech recognition parameter based on the category of the grammar.