Updating a voice template

• US 8,775,178 B2
• Filed: 10/27/2009
• Issued: 07/08/2014
• Est. Priority Date: 10/27/2008
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Miki, Yukari, Noguchi, Masami
• Primary Examiner(s): Godbold, Douglas
• Assistant Examiner(s): Villena, Mark
• Application Number:
  US12/606,206
• Publication Number:
  US 20100106501A1
• Time in Patent Office: 1,715 Days
• Field of Search: 704/250, 704/270, 704/246, 379/202.01
• US Class Current: 704/244
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 17/04 Training, enrolment or model building
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting

Abstract

Updating a voice template for recognizing a speaker on the basis of a voice uttered by the speaker is disclosed. Stored voice templates indicate distinctive characteristics of utterances from speakers. Distinctive characteristics are extracted for a specific speaker based on a voice message utterance received from that speaker. The distinctive characteristics are compared to the characteristics indicated by the stored voice templates to selected a template that matches within a predetermined threshold. The selected template is updated on the basis of the extracted characteristics.

Claims

1. An apparatus for updating a voice template for recognizing a speaker on the basis of a voice uttered by the speaker, the apparatus comprising:

a storage unit for storing a plurality of voice templates, each voice template indicating distinctive characteristics uttered by a corresponding one of a plurality of speakers;

a characteristics extraction unit for extracting distinctive characteristics uttered by a specific speaker on the basis of a voice message uttered by the specific speaker;

a selection unit for selecting a voice template from the plurality of voice templates stored in the storage unit on the basis of a degree of similarity between the distinctive characteristics indicated by the selected voice template and the distinctive characteristics extracted by the characteristics extraction unit;

an update unit for updating the selected voice template on the basis of the distinctive characteristics extracted by the characteristics extraction unit;

a voice recognition unit for performing voice recognition on the voice message uttered by the specific speaker;

a keyword extraction unit for extracting a keyword from specific data obtained as a result of the voice recognition by the voice recognition unit; and

a determination unit for determining a plurality of candidate templates on the basis of the keyword extracted by the keyword extraction unit, wherein the selected voice template is selected by the selection unit from among from the plurality of candidate templates determined by the determination unit.

2. The apparatus of claim 1, wherein the degree of similarity between the distinctive characteristics indicated by the selected voice template and the distinctive characteristics extracted by the characteristics extraction unit is greater than or equal to a predetermined threshold value.

3. The apparatus according to claim 1, further comprising:

an identification unit for identifying an addressee of the voice message uttered by the specific speaker; and

a determination unit for determining a plurality of candidate templates on the basis of a history of exchange of voice messages of the addressee identified by the identification unit;

wherein the selected voice template is selected by the selection unit from among from the plurality of candidate templates determined by the determination unit.

4. The apparatus according to claim 3, wherein each of the plurality of candidate templates indicates distinctive characteristics uttered by a corresponding one of a plurality of speakers with which the identified addressee has exchanged voice messages.

5. The apparatus according to claim 1, wherein each of the plurality of candidate templates indicates distinctive characteristics uttered by a corresponding one of a plurality of speakers related to the keyword extracted by the keyword extraction unit.

6. A method for updating a voice template for recognizing a speaker on the basis of a voice uttered by the speaker, the method comprising:

storing a plurality of voice templates, each voice template indicating distinctive characteristics uttered by a corresponding one of a plurality of speakers;

extracting distinctive characteristics uttered by a specific speaker on the basis of a voice message uttered by the specific speaker;

selecting a voice template from the stored plurality of voice templates on the basis of a degree of similarity between the distinctive characteristics indicated by the selected voice template and the extracted distinctive characteristics;

updating the selected specific voice template on the basis of the extracted distinctive characteristics;

performing voice recognition on the voice message uttered by the specific speaker;

extracting a keyword from specific data obtained as a result of the voice recognition; and

determining a plurality of candidate templates on the basis of the extracted keyword, and wherein selecting the voice template comprises selecting the voice template from among from the plurality of candidate templates.

7. The method of claim 6, wherein selecting a voice template comprises selecting a voice template so that the degree of similarity between the distinctive characteristics indicated by the selected voice template and the extracted distinctive characteristics is greater than or equal to a predetermined threshold value.

8. The method of claim 6, further comprising:

identifying an addressee of the voice message uttered by the specific speaker; and

determining a plurality of candidate templates on the basis of a history of exchange of voice messages of the addressee; and

wherein selecting the voice template comprises selecting the voice template from among from the plurality of candidate templates.

9. The method of claim 8, wherein each of the plurality of candidate templates indicates distinctive characteristics uttered by a corresponding one of a plurality of speakers with which the identified addressee has exchanged voice messages.

10. The method of claim 6, wherein each of the plurality of candidate templates indicates distinctive characteristics uttered by a corresponding one of a plurality of speakers related to the extracted keyword.

11. The method of claim 6, wherein the voice message uttered by the specific speaker is exchanged between the specific speaker and a speaker other than the specific speaker via a telephone line, and wherein extracting distinctive characteristics uttered by a specific speaker on the basis of a voice message uttered by the specific speaker comprises extracting the distinctive characteristics uttered by the specific speaker on the basis of the voice message obtained from a specific telephone line.

12. The method of claim 6, further comprising authenticating the specific speaker using the selected voice template causing the specific speaker to hear a voice message uttered by a speaker other than the specific speaker.

13. A computer program product for updating a voice template for recognizing a speaker on the basis of a voice uttered by the speaker, the computer program product comprising:

a computer readable storage medium device having computer readable program embodied therewith, the computer readable program code comprising;

computer readable program code configured to store a plurality of voice templates, each voice template indicating distinctive characteristics uttered by a corresponding one of a plurality of speakers;

computer readable program code configured to extract distinctive characteristics uttered by a specific speaker on the basis of a voice message uttered by the specific speaker;

computer readable program code configured to select a voice template from the stored plurality of voice templates on the basis of a degree of similarity between the distinctive characteristics indicated by the selected voice template and the extracted distinctive characteristics;

computer readable program code configured to update the selected specific voice template on the basis of the extracted distinctive characteristics;

computer readable program code configured to perform voice recognition on the voice message uttered by the specific speaker;

computer readable program code configured to extract a keyword from specific data obtained as a result of the voice recognition; and

computer readable program code configured to determine a plurality of candidate templates on the basis of the extracted keyword, and wherein computer readable program code configured to select a voice template comprises computer readable program code configured to select the voice template from among from the plurality of candidate templates.

14. The computer program product of claim 13, wherein computer readable program code configured to select a voice template comprises computer readable program code configured to select a voice template so that the degree of similarity between the distinctive characteristics indicated by the selected voice template and the extracted distinctive characteristics is greater than or equal to a predetermined threshold value.

15. The computer program product of claim 13, further comprising:

computer readable program code configured to identify an addressee of the voice message uttered by the specific speaker; and

computer readable program code configured to determine a plurality of candidate templates on the basis of a history of exchange of voice messages of the addressee; and

wherein computer readable program code configured to select a voice template comprises computer readable program code configured to select selecting the voice template from among from the plurality of candidate templates.

16. The computer program product of claim 15, wherein each of the plurality of candidate templates indicates distinctive characteristics uttered by a corresponding one of a plurality of speakers with which the identified addressee has exchanged voice messages.

17. The computer program product of claim 13, wherein each of the plurality of candidate templates indicates distinctive characteristics uttered by a corresponding one of a plurality of speakers related to the extracted keyword.