Speaker identification method, speaker identification apparatus, and information management method

• US 9,911,421 B2
• Filed: 06/05/2014
• Issued: 03/06/2018
• Est. Priority Date: 06/10/2013
• Status: Active Grant

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Inventors: Tsujikawa, Misaki, Banba, Yutaka
• Primary Examiner(s): ALBERTALLI, BRIAN LOUIS
• Application Number:
  US14/419,056
• Publication Number:
  US 20150194155A1
• Time in Patent Office: 1,370 Days
• CPC Class Codes: G06F 16/60 of audio data
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques
• CPC Class Codes: H04N 21/42203 sound input device, e.g. microphone
• CPC Class Codes: H04N 21/4415 using biometric characteristics of the user, e.g. by voice recognition or fingerprint scanning cryptography using biological data H04L9/3231; authentication in networks using biometric H04L63/0861
• CPC Class Codes: H04N 21/4532 involving end-user characteristics, e.g. viewer profile, preferences monitoring of user activities for profile generation for accessing a video database G06F16/739; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18

Abstract

The speaker identification system has a voice acquisition unit that acquires voice information of a speaker, and a database management unit that determines whether or not the speaker corresponding to the acquired voice information matches a speaker corresponding to registered voice information in connection with content information on a content, that acquires content information on the content displayed on a device at the time of acquisition of the voice information and stores the acquired content information in connection with the registered voice information in a case where it is determined that the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information, and that stores the acquired voice information in the database as registered voice information in a case where it is determined that the speaker corresponding to the acquired voice information does not match the speaker corresponding to the registered voice information.

Claims

1. A speaker identification method for identifying a speaker in the vicinity of a device displaying a content, the method comprising the steps of:

displaying a content on the device;

acquiring voice information of the speaker during display of the content on the device;

determining whether or not the speaker corresponding to the acquired voice information matches a speaker corresponding to registered voice information stored in a voice database in connection with content information on a content, the content information including a name of a cast member appearing in the content, and the registered voice information being voice information of a member belonging to a predetermined group;

in a case where it is determined that the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information stored in the voice database, acquiring the content information on the content displayed on the device from a content database for storing information to identify the content, and the content information in connection with each other at the time of acquisition of the voice information, and storing the acquired content information in connection with the registered voice information;

in a case where it is determined that the speaker corresponding to the acquired voice information does not match the speaker corresponding to the registered voice information stored in the voice database, compiling the acquired voice information in an internal memory which is different from the voice database without updating the voice database at the time of the determining;

identifying a speaker of a plurality of pieces of voice information compiled in the internal memory for a predetermined period of time, and extracting, among the plurality of pieces of voice information, two or more pieces of voice information identified as corresponding to a same speaker;

counting the number of the extracted pieces of voice information by the same speaker, and storing one among the extracted pieces of voice information in the voice database as registered voice information of a new member belonging to the predetermined group in a case where the counted number indicates a predetermined number or more; and

in a case where it is determined that the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information stored in the voice database, extracting the name of the cast member from the content information linked to the speaker corresponding to the acquired voice information, and referencing a service database in which names of cast members are stored in connection with services to be provided to speakers, thereby specifying a service associated with the name of the cast member as a candidate for a service to be provided.

2. The speaker identification method according to claim 1, wherein the content information includes a name of the content and a name of a person associated with the content.

3. The speaker identification method according to claim 1, further comprising the step of classifying a plurality of contents associated with the registered voice information into a plurality of genres, calculating, for each of the plurality of genres, a percentage of contents classified into each of the genres from among the plurality of contents, and storing the percentage of contents calculated for each of the plurality of genres in the voice database in connection with the registered voice information.

4. The speaker identification method according to claim 1, wherein

the voice database stores content information in connection with a service to be provided to a speaker who views a content corresponding to the content information, and

the method further comprises the step of, in a case where it is determined that the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information stored in the voice database, specifying the content information stored in connection with the registered voice information, specifying a service associated with the specified content information, and providing the specified service to the speaker.

5. The speaker identification method according to claim 4, further comprising the steps of:

determining whether at least one available service exists or not and whether or not the at least one available service is provided at a predetermined service providing timing; and

in a case where it is determined that the at least one available service exists and that the at least one available service is provided at the predetermined service providing timing, displaying candidates for the at least one available service on the device.

6. The speaker identification method according to claim 5, further comprising the steps of:

providing the speaker with a service that is selected by the speaker from among the displayed candidates for the at least one available service; and

storing the provided service in the voice database in connection with the registered voice information.

7. The speaker identification method according to claim 4, wherein the service includes a service for distributing a content to be displayed on the device, or a service for distributing an advertisement to be displayed on the device.

8. The speaker identification method according to claim 1, wherein, in a case where partial voice information of the registered voice information registered in the voice database is not acquired for a predetermined period of time or longer, the partial voice information and/or information associated with the partial voice information is deleted from the voice database.

9. The speaker identification method according to claim 1, wherein whether or not the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information stored in the voice database in connection with the content information on the content is determined by extracting text information from the acquired voice information and analyzing spoken words based on the extracted text information.

10. The speaker identification method according to claim 5, wherein

the device includes a television, and

the service providing timing is a timing at which an operation for displaying a program guide for the television is executed.

11. A speaker identification apparatus for identifying a speaker, comprising:

a display unit that displays a content;

a voice acquisition unit that acquires voice information of a speaker in the vicinity of the speaker identification apparatus during display of the content on the display unit;

a voice database for storing registered voice information in connection with content information on a content, the content information including a name of a cast member appearing in the content, and the registered voice information being voice information of a member belonging to a predetermined group;

a content database for storing information to identify the content, and the content information in connection with each other;

a determination unit that determines whether or not the speaker corresponding to the voice information acquired by the voice acquisition unit matches a speaker corresponding to the registered voice information stored in the voice database in connection with the content information;

a database update unit that acquires the content information on the content displayed on the display unit from the content database at the time of acquisition of the voice information and stores the acquired content information in connection with the registered voice information, in a case where the determination unit determines that the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information stored in the voice database;

a compilation unit that compiles the acquired voice information in an internal memory which is different from the voice database in a case where the determination unit determines that the speaker corresponding to the acquired voice information does not match the speaker corresponding to the registered voice information stored in the voice database, without updating the voice database at the time of the determination by the determining unit;

an extraction unit that identifies a speaker of a plurality of pieces of voice information compiled in the memory for a predetermined of time, and extracts, among the plurality of pieces of voice information, two or more pieces of voice information identified as corresponding to a same speaker;

a database storage unit that counts the number of the extracted pieces of voice information by the same speaker, and stores one among the extracted pieces of voice information in the voice database as registered voice information of a new member belonging to the predetermined group in a case where the counted number indicates a predetermined number or more; and

a specification unit that extracts the name of the cast member from the content information linked to the speaker corresponding to the acquired voice information, and references a service database in which names of cast members are stored in connection with services to be provided to speakers, thereby specifying a service associated with the name of the cast member as a candidate for a service to be provided.

12. An information management method of a speaker identification system for identifying a speaker in the vicinity of a device displaying a content, the method comprising the steps of:

displaying a content on the device;

receiving voice information of the speaker, the voice information being acquired during display of the content on the device;

determining whether or not the speaker corresponding to the received voice information matches a speaker corresponding to registered voice information stored in a voice database in connection with content information on a content, the content information including a name of a cast member appearing in the content, and the registered voice information being voice information of a member belonging to a predetermined group;

in a case where it is determined that the speaker corresponding to the received voice information matches the speaker corresponding to the registered voice information stored in the voice database, acquiring the content information on the content displayed on the device from a content database for storing information to identify the content, and the content information in connection with each other at the time of acquisition of the voice information, and storing the received content information in connection with the registered voice information;

in a case where it is determined that the speaker corresponding to the received voice information does not match the speaker corresponding to the registered voice information stored in the voice database, compiling the acquired voice information in an internal memory which is different from the voice database without updating the voice database at the time of the determining;

identifying a speaker of a plurality of pieces of voice information compiled in the memory for a predetermined period of time, and extracting, among the plurality of pieces of information, two or more pieces of voice information identified as corresponding to a same speaker;

counting the number of the extracted pieces of voice information by the same speaker, and storing one among the extracted pieces of voice information in the voice database as registered voice information of a new member belonging to the predetermined group in a case where the counted number indicates a predetermined number or more; and

in a case where it is determined that the speaker corresponding to the acquired voice information matches the speaker corresponding to the registered voice information stored in the voice database, extracting the name of the cast member from the content information linked to the speaker corresponding to the acquired voice information, and referencing a service database in which names of cast members are stored in connection with services to be provided to speakers, thereby specifying a service associated with the name of the cast member as a candidate for a service to be provided.