Voice command processing system and computer therefor, and voice command processing method

• US 7,299,187 B2
• Filed: 02/10/2003
• Issued: 11/20/2007
• Est. Priority Date: 02/13/2002
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Tahara, Yoshinori, Tomoda, Daisuke, Atake, Yoshinori, Mitsubo, Kikuo
• Primary Examiner(s): Hudspeth; David
• Assistant Examiner(s): Rider; Justin W.
• Application Number:
  US10/361,547
• Publication Number:
  US 20030154077A1
• Time in Patent Office: 1,744 Days
• Field of Search: 704/238, 704/239, 704/257, 704/275
• US Class Current: 704/275
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates

Abstract

When a user issued voice command does not match grammars registered in advance, the voice command is identified as a sentence (step S305). This sentence is compared with the registered grammars to calculate a similarity (step S307). When the similarity is higher than a first threshold value (TH1), the voice command is executed (step S315). When the similarity is equal to or lower than the first threshold value (TH1) and higher than a second threshold value (TH2), command choices are displayed for the user and the user is permitted to select a command to be executed (step S319). When the similarity is equal to or lower than the second threshold value (TH2), the command is not executed (step S321). Furthermore, once a command has been executed it is added as a grammar, so that it can be identified when next it is used.

Claims

1. A voice command processing system, for processing a user issued voice command, comprising:

registration means for the advance registration of multiple executable commands;

acquisition means for obtaining said user issued voice command;

determination means for determining whether one of said executable commands registered in said registration means matches said user issued voice command obtained by said acquisition means;

execution means for executing one of said registered commands if performing a process to execute said user issued voice command matches one of said grammars;

calculation means for, when said determination means ascertains that said voice command does not match any of said registered executable commands, analyzing said voice issued command as a dictated sentence, and calculating for said registered commands and said sentence a similarity, wherein said similarity is based on a comparison of a reading of words in said sentence analyzed with a reading of words in said registered executable command; and

request means for when there are multiple registered executable commands for which the similarities fall above a first threshold, notifying said user of said registered executable commands, and requesting that said user determine whether said registered commands for which the similarities fall above said first threshold should be executed.

2. The voice command processing system according to claim 1, further comprising:

execution instruction means for instructing the execution of a registered executable command for which said similarity obtained by said calculation means falls within a predetermined range.

3. The voice command processing system according to claim 1, further comprising:

request means for, when there is one or multiple registered commands for which similarities fall between the first threshold and a second threshold lower than the first threshold, notifying said user of said registered commands for which similarities fall between the first and the second threshold, and requesting that said user determine whether said registered commands for which similarities fall between the first and the second threshold should be executed.

4. A computer, for executing user issued voice commands, comprising:

a grammar registration unit for registering one or more commands executable by said computer as grammars;

a command processor for accepting a user issued voice command and for executing one of said registered commands if said user issued voice command matches one of said grammars;

a voice recognition engine for analyzing, as a dictated sentence, said user issued voice command accepted by said command processor when it is impossible for said accepted voice command to be matched to one of said grammars;

a voice command similarity calculator for comparing a reading of words in said sentence analyzed by said voice recognition engine to a reading of words in each of said commands registered in said grammar registration unit, and calculating for said voice command a similarity to each of said registered commands; and

a score data registration unit for registering words having a reading similar to the reading of the words in each of said registered commands and a score for each of said similar reading words based on a similarity of the reading of said words in said registered command and said similar reading words;

wherein said voice command similarity calculator employs said scores when calculating said voice command similarity.

5. The computer according to claim 4, wherein said voice command processor executes said registered command when said similarity exceeds a predetermined threshold value.

6. A voice command processing method that permits a computer to perform a user issued voice command comprising the steps of:

accepting a user issued voice command;

determining whether said voice command can be identified as a previously registered command for an operation;

analyzing said voice command as a dictated sentence when it is impossible for said accepted voice command to be identified as a previously registered command for an operation;

calculating a score for each word in said sentence based on a similarity of the reading of said each word in said sentence and words in a previously registered command for an operation;

employing said scores to calculate a similarity for said sentence relative to said previously registered command; and

registering, as a new registered command for said operation and in correlation with said sentence, said accepted voice command if said similarity falls within a predetermined range.

7. The voice command processing method according to claim 6, further comprising the step of:

when said similarity falls within said predetermined range, executing said registered command for which said similarity has been calculated.

8. The voice command processing method according to claim 6, further comprising the steps of:

at least either when there are two or more registered commands for which similarities fall above a first threshold, or when there are one or more registered commands for which said similarities fall between said first threshold and a second threshold, set lower than said first threshold, displaying for said user said registered commands for which similarities have been obtained, and requesting that said user select a registered command to be executed.

9. A voice command processing method that permits a computer to execute a user issued voice command comprising the steps of:

determining that the user issued voice command cannot be identified as an executable command registered (W2) in advance;

analyzing as a dictated sentence said user issued voice command;

comparing said analyzed user issued voice command (W1) with the registered command (W2) to calculate a similarity S(W1,W2);

if said similarity S(W1,W2) is higher than a first threshold value (TH1), executing the registered command (W2); and

if said similarity S(W1,W2) is lower than said first threshold value (TH1) and higher than a second threshold value (TH2) lower than said first threshold value (TH1), displaying said registered command (W2) for said user and requesting said user determine whether said registered command (W2) should be executed,

whereby said similarity S(W1,W2) is represented by a value obtained by dividing, by the number (Vn) of applicable words for the calculation of said similarity S(W1,W2), a sum of scores (s) that is based on a similarity of a reading of the i-th word (w1(i)) of said voice command and the j-th word (w2(j)) of said registered command (W2).

10. A machine readable storage, having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to perform the steps of:

accepting a user issued voice command;

determining whether said voice command can be identified as a previously registered command for an operation;

analyzing said voice command as a dictated sentence when it is impossible for said accepted voice command to be identified as a previously registered command for an operation;

calculating a score for each word in said sentence based on a similarity of the reading of said each word in said sentence and words in a previously registered command for an operation;

employing said scores to calculate a similarity for said sentence relative to said previously registered command; and

registering, as a new registered command for said operation and in correlation with said sentence, said accepted voice command if said similarity falls within a predetermined range.

11. The machine readable storage of claim 10, further causing the machine to perform the step of:

when said similarity falls within said predetermined range, executing said registered command for which said similarity has been calculated.

12. The machine readable storage of claim 10, further causing the machine to perform the steps of:

at least either when there are two or more registered commands for which similarities fall above a first threshold, or when there are one or more registered commands for which said similarities fall between said first threshold and a second threshold, set lower than said first threshold, displaying for said user said registered commands for which similarities have been obtained, and requesting that said user select a registered command to be executed.

13. A machine readable storage, having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to perform the steps of:

determining that the user issued voice command cannot be identified as an executable command registered (W2) in advance;

analyzing as a dictated sentence said user issued voice command; and

comparing said analyzed user issued voice command (W1) with the registered command (W2) to calculate a similarity S(W1,W2);

if said similarity S(W1,W2) is higher than a first threshold value (TH1), executing the registered command (W2); and

if said similarity S(W1,W2) is lower than said first threshold value (TH1) and higher than a second threshold value (TH2) lower than said first threshold value (TH1), displaying said registered command (W2) for said user and requesting said user determine whether said registered command (W2) should be executed,

whereby said similarity S(W1,W2) is represented by a value obtained by dividing, by the number (Vn) of applicable words for the calculation of said similarity S(W1,W2), the sum of scores (s) that is based on a similarity of a reading of the i-th word (w1(i)) of said voice command and the j-th word (w2(j)) of said registered command (W2).