METHOD AND SYSTEM FOR SHARING PORTABLE VOICE PROFILES

• US 20120284027A1
• Filed: 06/14/2012
• Published: 11/08/2012
• Est. Priority Date: 09/28/2006
• Status: Active Grant

• Current Assignee: reQall, Inc.
• Original Assignee: reQall, Inc.
• Inventors: Mallett, Jacqueline, Vemuri, Sunil, Machiraju, N. Rao
• Granted Patent: US 8,990,077 B2
• Time in Patent Office: Days
• US Class Current: 704/246
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 17/00 Speaker identification or verification techniques

Abstract

An embodiment of the present invention provides a speech recognition engine that utilizes portable voice profiles for converting recorded speech to text. Each portable voice profile includes speaker-dependent data, and is configured to be accessible to a plurality of speech recognition engines through a common interface. A voice profile manager receives the portable voice profiles from other users who have agreed to share their voice profiles. The speech recognition engine includes speaker identification logic to dynamically select a particular portable voice profile, in real-time, from a group of portable voice profiles. The speaker-dependent data included with the portable voice profile enhances the accuracy with which speech recognition engines recognize spoken words in recorded speech from a speaker associated with a portable voice profile.

Claims

1. A system comprising:

a storage device configured to store portable voice profiles of a plurality of speakers; and

a computer-based system communicatively coupled to the storage device, the computer-based system including (1) a recognition engine configured to identify individual ones of the speakers from the respective portable voice profiles, wherein each portable voice profile is associated with and customized to a particular one of the speakers, using speaker identification logic to analyze voice signals to identify the particular speaker and then dynamically select a particular portable voice profile associated with the identified speaker;

(2) a local group manager configured to manage access privileges of a user according to said identification.

2. The system of claim 1, wherein a group is initially defined by an administrative application executing on a remote system, and the local group manager provides a listing of all groups the user is allowed to join.

3. The system of claim 1, wherein the local group manager is further configured to authenticate the user according to said identification.

4. The system of claim 1, wherein the local group manager is further configured to authenticate the user using the selected portable voce profile.

5. A system, comprising:

a storage device configured to store a plurality of portable voice profiles; and

a computer-based system communicatively coupled to the storage device, the computer-based system including (1) a recognition engine including speaker identification logic configured to analyze a voice signal to identify a speaker through a particular portable voice profile associated with the identified speaker; and

(2) a local voice profile manager configured to receive the plurality of portable voice profiles, each portable voice profile associated with a speaker and including speaker-dependent data accessible to a plurality of recognition engines through an interface, the speaker-dependent data to enhance an accuracy with which each recognition engine in the plurality of recognition engines recognizes spoken words in a voice signal from a speaker associated with a portable voice profile.

6. The system of claim 5, wherein the system is configured to facilitate establishing a user of the system as a member of a group and to facilitate establishing connections with other users as members of the group.

7. The apparatus of claim 5, wherein the system is configured to facilitate establishing access rights of a user using the portable voice profiles.

8. The system of claim 5, wherein the speaker-dependent data of each portable voice profile includes acoustic characteristics of a speaker associated with the portable voice profile.

9. The system of claim 5, wherein the speaker-dependent data of each portable voice profile is derived from feedback provided by any one of said plurality of recognition engines.

10. The system of claim 5, wherein the speaker-dependent data of each portable voice profile is derived without explicit training and is updated over time.

11. The system of claim 5, wherein the local voice profile manager is further configured to authenticate the user according to said identification.

12. The system of claim 9, wherein the local voice profile manager is further configured to authenticate the user via the particular portable voice profile associated with the user.

13. A computer-implemented method comprising:

receiving a voice signal from a speaker using a recognition engine executing on a computer-based platform;

analyzing the voice signal to identify the speaker through use of a plurality of voice profiles, each associated with a respective speaker;

establishing access control privileges for the speaker identified via the portable voice profiles, wherein at least one of the portable voice profiles includes data derived from use with a variety of recognition engines.

14. The method of claim 13, wherein the portable voice profiles are exchanged over a network, and each portable voice profile is associated with a speaker and includes speaker-dependent data accessible to a plurality of recognition engines through an interface, the speaker-dependent data to enhance an accuracy with which each recognition engine in the plurality of recognition engines recognizes spoken words in recorded speech from a respective speaker associated with a respective portable voice profile.

15. The method of claim 13, further comprising authenticating the speaker using the identity.