Method for the training or adaptation of a speech recognition device

• US 7,050,550 B2
• Filed: 05/10/2002
• Issued: 05/23/2006
• Est. Priority Date: 05/11/2001
• Status: Expired due to Term

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Steinbiss, Volker, Eisele, Thomas
• Primary Examiner(s): Escalante, Ovidio
• Application Number:
  US10/143,641
• Publication Number:
  US 20030012347A1
• Time in Patent Office: 1,474 Days
• Field of Search: 379/88.03, 379/88.01, 379/88.04, 379/355.01, 379/88.16, 455/563, 455/564, 455/569.1, 704/200, 704/231, 704/246, 704/247, 704/270, 704/275, 704/258
• US Class Current: 379/88.01
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

The invention relates to a method for the training or adaptation of a speech recognition device used to act upon functions of an electrical appliance, for example the triggering of a voice dial in a mobile telephone terminal. In order to structure the training and/or adaptation of the speech recognition device to improve user comfort, a method is proposed with the following steps:

• • performance of a speech input;
  • processing of the speech input by means of the speech recognition device for the production of a speech recognition result;
  • if the speech recognition result can be allocated to a function of the electrical appliance, action upon the allocatable function of the electrical appliance;
  • training or adaptation of the speech recognition device on the basis of the speech recognition result associated with the speech input made, if the action upon the allocatable function of the electrical appliance does not cause a user input expressing rejection.

Claims

1. A method for the training or adaptation of a speech recognition device which is used to act upon functions of an electrical appliance, in which the following steps are provided:

performance of a speech input;

processing of the speech input by means of the speech recognition device;

producing a speech recognition result;

if the speech recognition result can be allocated to a function of the electrical appliance, action upon the allocatable function of the electrical appliance;

training or adaptation of the speech recognition device on the basis of the speech recognition result associated with the speech input made, if the action upon the allocatable function of the electrical appliance does not cause a user input expressing rejection, characterized in that, if after a speech input a user input is made expressing the rejection of a speech recognition result and if, subsequently, a user input is made by means of another input method, it is provided that, taking into account the input made by means of the other input method, a reliability grade for the speech recognition result is determined and a training operation or adaptation of the speech recognition device performed, provided that the speech input and the user input made by means of the other input method represent the same word or word sequence.

2. A method as claimed in claim 1, characterized in that the training or adaptation of the speech recognition device takes place if the performance of the function of the electrical appliance is not interrupted by a user input.

3. A method as claimed in claim 1, characterized in that, for a speech recognition result of the speech recognition device, a reliability grade is determined, and that a training or adaptation of the speech recognition device is carried out on the basis of the speech recognition result corresponding to the speech input made if the reliability grade of the speech recognition result falls within a prescribed target range.

4. A method as claimed in claim 1, characterized in that,

for a speech recognition result of the speech recognition device, an initial reliability grade is determined,

in that the training or adaptation of the speech recognition device is performed on the basis of the speech recognition result corresponding to the speech input made if the initial reliability grade of the speech recognition result falls within a prescribed initial target range,

in that if the initial reliability grade does not fall within the initial target range and also if the subsequent user input is made by means of another input method, it is provided that, taking into account the input made by means of the other input method, a second reliability grade for the speech recognition result is determined and a training operation or adaptation is performed on the basis of the speech recognition result taking into account the user input made by means of the other input method if the second reliability grade falls within a second prescribed target range.

5. A method as claimed in claim 1, characterized in that

the speech recognition device also supplies in each case at least one speech recognition result alternative,

in that the training or adaptation of the speech recognition device is carried out on the basis of the speech recognition result associated with the latest speech input, if the action upon the function of the electrical appliance allocatable to the speech recognition result does not cause a user input expressing rejection of the speech recognition result, and if

the speech recognition result corresponds to a speech recognition result alternative to the speech input before last and

a user input expressing rejection of the speech recognition result before last is made,

a further training operation or further adaptation of the speech recognition device takes place provided that the latest speech recognition result and the one before last represent the same word or the same word sequence.

6. A method as claimed in claim 1, characterized in that an acoustic or visual reproduction of a speech recognition result is provided.

7. A method as claimed in claim 1, characterized in that the electrical appliance is a mobile telephone.

8. A method as claimed in claim 7, characterized in that, by means of a speech input, a dialing process is triggered.

9. An electrical appliance with a speech recognition device, used to act upon functions of the electrical appliance, in which the electrical appliance and the speech recognition device are designed in such a way that the following steps are used for the training or adaptation of the speech recognition device:

performance of a speech input;

processing of the speech input by means of the speech recognition device;

producing a speech recognition result;

if the speech recognition result can be allocated to a function of the electrical appliance, action upon the allocatable function of the electrical appliance;

training or adaptation of the speech recognition device on the basis of the speech recognition result associated with the speech input made, if the action upon the allocatable function of the electrical appliance does not cause a user input expressing rejection,;

characterized in that, if after a speech input a user input is made expressing the rejection of a speech recognition result and if subsequently, a user input is made by means of another input method, it is provided that, taking into account the input made by means of the other input method, a reliability grade for the speech recognition result is determined and a training operation or adaptation of the speech recognition device performed, provided that the speech input and the user input trade by means of the other input method represent the same word or word sequence.

10. An electrical appliance as claimed in claim 9, characterized in that, for a speech recognition result of the speech recognition device, a reliability grade is determined, and that a training or adaptation of the speech recognition device is carried out on the basis of the speech recognition result corresponding to the speech input made if the reliability grade of the speech recognition result falls within a prescribed target range.

11. An electrical appliance as claimed in claim 9, characterized in that,

for a speech recognition result of the speech recognition device, an initial reliability grade is determined,

in that the training or adaptation of the speech recognition device is performed on the basis of the speech recognition result corresponding to the speech input made if the initial reliability grade of the speech recognition result falls within a prescribed initial target range,

in that if the initial reliability grade does not fall within the initial target range and also if the subsequent user input is made by means of another input method, it is provided that, taking into account the input made by means of the other input method, a second reliability grade for the speech recognition result is determined and a training operation or adaptation is performed on the basis of the speech recognition result taking into account the user input made by means of the other input method if the second reliability grade falls within a second prescribed target range.

12. An electrical appliance as claimed in claim 9, characterized in that

the speech recognition device also supplies in each case at least one speech recognition result alternative,

in that the training or adaptation of the speech recognition device is carried out on the basis of the speech recognition result associated with the latest speech input, if the action upon the function of the electrical appliance allocatable to the speech recognition result does not cause a user input expressing rejection of the speech recognition result, and if

the speech recognition result corresponds to a speech recognition result alternative to the speech input before last and

a user input expressing rejection of the speech recognition result before last is made,

a further training operation or further adaptation of the speech recognition device takes place provided that the latest speech recognition result and the one before last represent the same word or the same word sequence.

13. A mobile terminal with a speech recognition device used to act upon functions of the mobile telephone terminal, in which the mobile telephone terminal and the speech recognition device are designed in such a way that the following steps are used for the training or adaptation of the speech recognition device:

performance of a speech input;

processing of the speech input by means of the speech recognition device (4);

production of a speech recognition result;

if the speech recognition result can be allocated to a function of the mobile telephone terminal, action upon the allocatable function of the mobile telephone terminal;

training or adaptation on the basis of the speech recognition result associated with the speech input made, if the action upon the allocatable function of the mobile telephone terminal does not cause a user input expressing rejection,

characterized in that, if after a speech input a user input is made expressing the rejection of a speech recognition result and if, subsequently, a user input is made by means of another input method, it is provided that, taking into account the input made by means of the other input method, a reliability grade for the speech recognition result is determined and a training operation or adaptation of the speech recognition device performed, provided that the speech input and the user input made by means of the other input method represent the same word or word sequence.

14. A mobile terminal as claimed in claim 13, characterized in that the training or adaptation of the speech recognition device takes place if the performance of the function of the electrical appliance is not interrupted by a user input.

15. A mobile terminal as claimed in claim 13, characterized in that, for a speech recognition result of the speech recognition device, a reliability grade is determined, and that a training or adaptation of the speech recognition device is carried out on the basis of the speech recognition result corresponding to the speech input made if the reliability grade of the speech recognition result falls within a prescribed target range.

16. A mobile terminal as claimed in claim 13, characterized in that,

for a speech recognition result of the speech recognition device, an initial reliability grade is determined,

in that the training or adaptation of the speech recognition device is performed on the basis of the speech recognition result corresponding to the speech input made if the initial reliability grade of the speech recognition result falls within a prescribed initial target range,

in that if the initial reliability grade does not fall within the initial target range and also if the subsequent user input is made by means of another input method, it is provided that, taking into account the input made by means of the other input method, a second reliability grade for the speech recognition result is determined and a training operation or adaptation is performed on the basis of the speech recognition result taking into account the user input made by means of the other input method if the second reliability grade falls within a second prescribed target range.

17. A mobile terminal as claimed in claim 13, characterized in that

the speech recognition device also supplies in each case at least one speech recognition result alternative,

in that the training or adaptation of the speech recognition device is carded out on the basis of the speech recognition result associated with the latest speech input, if the action upon the function of the electrical appliance allocatable to the speech recognition result does not cause a user input expressing rejection of the speech recognition result, and if

the speech recognition result corresponds to a speech recognition result alternative to the speech input before last and

a user input expressing rejection of the speech recognition result before last is made,

a further training operation or further adaptation of the speech recognition device takes place provided that the latest speech recognition result and the one before last represent the same word or the same word sequence.