Voice print recognition software system for voice identification and matching

• US 8,595,007 B2
• Filed: 06/05/2007
• Issued: 11/26/2013
• Est. Priority Date: 06/15/2006
• Status: Active Grant

• Current Assignee: NITV Federal Services LLC
• Original Assignee: NITV Federal Services LLC
• Inventors: Kane, James A.
• Primary Examiner(s): Godbold, Douglas
• Assistant Examiner(s): Villena, Mark
• Application Number:
  US11/758,474
• Publication Number:
  US 20080091425A1
• Time in Patent Office: 2,366 Days
• Field of Search: 704/247, 704/246, 704/500, 704/273, 455/411
• US Class Current: 704/246
• CPC Class Codes: G10L 17/04 Training, enrolment or model building

Abstract

Positive identification of local inhabitants plays an important role in modern military, police and security operations. Since terrorists use all means to masquerade as local inhabitants, the identification of terrorist or hostile suspects becomes an increasingly complicated task. The instant software solution will assist military, police and security forces in the identification of suspects using Voice Print Recognition (VPR) technology. Our VPR software will compare and recognize, or match, specific voice samples with stored, digital voice models (voice prints) for the purpose of establishing or verifying identity. VPR software will support an operator'"'"'s decision and situational awareness through the verification of a person'"'"'s identity (for instance: remote access control), but more importantly will assist in the identification of suspect individuals (identifying suspects among a large group of captured individuals). This second application is critical for the modern counter and anti-terrorist operations environment. The VPR system will be easy to use, fast, and helpful to users with minimal operational training. The VPR system will provide a method, as practiced on or via access to a computing device, which includes software for acquisition of voice records, storage of such records, identification algorithms, user and software interfaces. The system will also have server and client applications in its implementation.

Claims

1. A method for identification verification comprising the steps of:

employing a computer based system for obtaining at least one first spoken voice sample of an individual speaking at least one piece of specific data defined as a plurality of scripted words, phrases, answers to predetermined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

obtaining a first spoken voice print comprised of one or more vocal characteristics obtained from said voice sample, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple languages;

storing said at least one first spoken voice print of said individual in a database of said computer based system;

obtaining a voice sample from a cooperative or non-cooperative test subject to be verified or identified as an individual within said database;

employing said computer based system for obtaining at least one second spoken voice print from said voice sample from said cooperative or non-cooperative test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or answers to predetermined questions and at least one piece of non-specific data defined as one or more non-predetermined words or phrases;

said computer based system using a first algorithm for determining one or more characteristics of said spoken voice print if said test subject is a cooperative individual and a second algorithm for determining one or more characteristics of said second spoken voice print if said test subject is a non-cooperative individual whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple languages;

comparing by said computer based system or a second remote computer system said at least one second spoken voice print of said test subject to said at least one first voice print of said individual within said database;

determining or verifying if said test subject is said individual within said database based on the results of said comparison; and

storing said at least one spoken voice print of said test subject in said database when it is determined that said test subject is not said individual within said database.

2. The method of claim 1, further comprising the step of employing means for obtaining at least one third spoken voice print of said test subject speaking at least one piece of structured data defined as a plurality of scripted words or phrases and at least one piece of unstructured data defined as one or more unscripted words or phrases;

comparing said at least one third spoken voice print of said test subject to said at least one second voice print of said test subject; and

determining if said test subject is said individual based on the results of said comparison.

3. The method of claim 2, wherein said test subject is a non-cooperative suspect.

4. The method of claim 3, wherein the step of obtaining said at least one second spoken voice print of said test subject is performed on site.

5. The method of claim 3, wherein the step of obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

6. The method of claim 3, wherein the step of obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

7. The method of claim 1, wherein the step of obtaining said at least one second spoken voice print of said test subject is performed on site.

8. The method of claim 1, wherein the step of obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

9. The method of claim 1, wherein the step of obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

10. A method for identification verification comprising the steps of:

obtaining at least one first spoken voice sample of an individual speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to predetermined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

obtaining a first spoken voice print comprised of one or more vocal characteristics obtained from said voice sample, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple language voice samples;

storing said at least one first spoken voice print of said individual in a database of a computer based system;

obtaining a voice sample from a cooperative or non-cooperative test subject to be verified or identified by voice analysis using a computer based system;

obtaining at least one second spoken voice print from said voice sample from said cooperative or non-cooperative test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to predetermined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases using a computer based system;

said computer based system using a first algorithm for determining one or more characteristics of said spoken voice print if said test subject is a cooperative individual and a second algorithm for determining one or more characteristics of said second spoken voice print if said test subject is a non-cooperative individual, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple language voice samples;

comparing said at least one second spoken voice print of said test subject to said at least one first voice print of said individual;

determining or verifying if said test subject is said individual based on the results of said comparison;

storing said at least one spoken voice print of said test subject in said database when it is determined that said test subject is not said individual.

11. The method of claim 10, further comprising the step of employing means for obtaining at least one third spoken voice print of a test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to predetermined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

comparing said at least one third spoken voice print of said test subject to said at least one second voice print of said test subject; and

determining if said test subject is said individual based on the results of said comparison.

12. The method of claim 11, wherein said test subject is a non-cooperative suspect.

13. The method of claim 12, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed on site.

14. The method of claim 12, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

15. The method of claim 12, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

16. The method of claim 10, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed on site.

17. The method of claim 10, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

18. The method of claim 10, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

19. A computer system implementing a method for identification verification comprising the steps of:

employing means for obtaining at least one first spoken voice sample of an individual speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to predetermined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

obtaining a first spoken voice print comprised of one or more vocal characteristics obtained from said voice sample, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple languages;

storing said at least one first spoken voice print of said individual in a database;

obtaining a voice sample from a cooperative or non-cooperative test subject to be verified or identified;

employing means for obtaining at least one second spoken voice print from said voice sample from said cooperative or non-cooperative test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to pre-determined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

said employing means using a first algorithm for determining one or more characteristics of said spoken voice print if said test subject is a cooperative individual and a second algorithm for determining one or more characteristics of said second spoken voice print if said test subject is a non-cooperative individual, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple languages;

comparing said at least one spoken voice print of said test subject to said at least one first voice print of said individual;

determining or verifying if said test subject is said individual based on the results of said comparison; and

storing said at least one spoken voice print of said test subject in said database if it is determined that said test subject is not said individual.

20. The computer system of claim 19, further comprising the step of employing means for obtaining at least one third spoken voice print of a test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to pre-determined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

comparing said at least one third spoken voice print of said test subject to said at least one second voice print of said test subject; and

determining if said test subject is said individual based on the results of said comparison.

21. The computer system of claim 19, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed on site.

22. The computer system of claim 19, wherein the step of obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

23. The computer system of claim 19, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

24. The computer system of claim 19, wherein said test subject is a non-cooperative suspect.

25. The computer system of claim 24, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed on site.

26. The computer system of claim 24, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

27. The computer system of claim 24, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

28. A non-transitory machine readable medium storing instructions, that when executed by a computer causes the computer to perform a set of operations comprising:

employing means for obtaining at least one first spoken voice sample of an individual speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to pre-determined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

obtaining a first spoken voice print comprised of one or more vocal characteristics obtained from said voice sample, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple languages;

storing said at least one first spoken voice print of said individual in a database;

obtaining a voice sample from a cooperative or non-cooperative test subject to be verified or identified by voice analysis;

employing means for obtaining at least one second spoken voice print from said voice sample from said cooperative or non-cooperative test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to pre-determined questions and at least one piece of non-specific data defined as one or more unscripted words or phrases;

said employing means using a first algorithm for determining one or more characteristics of said spoken voice print if said test subject is a cooperative individual and a second algorithm for determining one or more characteristics of said second spoken voice print if said test subject is a non-cooperative individual, whereby obtaining said voice print is not dependant upon a specific language spoken thereby allowing for voice prints being obtainable for multiple languages;

comparing said at least one spoken voice print of said test subject to said at least one first voice print of said individual;

determining or verifying if said test subject is said individual based on the results of said comparison; and

storing said at least one spoken voice print of said test subject in said database if it is determined that said test subject is not said individual.

29. The non-transitory machine readable medium of claim 28, further comprising the step of employing means for obtaining at least one third spoken voice print of a test subject speaking at least one piece of specific data defined as a plurality of scripted words, phrases, or responses to pre-determined questions and at least one piece of non-structured data defined as one or more unscripted words or phrases;

comparing said at least one third spoken voice print of said test subject to said at least one second voice print of said user; and

determining if said user is said individual based on the results of said comparison.

30. The non-transitory machine readable medium of claim 29, wherein said test subject is a non-cooperative suspect.

31. The non-transitory machine readable medium of claim 30, wherein the step of obtaining said at least one second spoken voice print of said test subject is performed on site.

32. The non-transitory machine readable medium of claim 30, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

33. The non-transitory machine readable medium of claim 30, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.

34. The non-transitory machine readable medium of claim 28, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed on site.

35. The non-transitory machine readable medium of claim 28, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject is performed from an intercepted communication.

36. The non-transitory machine readable medium of claim 28, wherein the step of employing means for obtaining said at least one second spoken voice print of said test subject occurs in a multi-speaker environment.