Multimodal biometric analysis

• US 7,362,884 B2
• Filed: 03/17/2006
• Issued: 04/22/2008
• Est. Priority Date: 03/17/2005
• Status: Active Grant

• Current Assignee: Tech5 USA Incorporated
• Original Assignee: Imageware Systems Incorporated
• Inventors: Dlab, Daniel, Willis, William Frederic, Lau, Johann Herbert
• Primary Examiner(s): Mehta; Bhavesh M.
• Assistant Examiner(s): KIM, CHONG R
• Application Number:
  US11/378,015
• Publication Number:
  US 20060210119A1
• Time in Patent Office: 767 Days
• Field of Search: None
• US Class Current: 382/116
• CPC Class Codes: G06F 18/2415 based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition
• CPC Class Codes: G06V 40/70 Multimodal biometrics, e.g. combining information from different biometric modalities

Abstract

A biometric system that uses readings from a plurality of biometrics of a user is disclosed. The biometric system includes a first and second biometric readers, a first and second biometric matching engines and a processor. The first biometric reader producing a first measured biometric that is processed by the first biometric matching engine to deliver a first value, which is indicative of a likelihood that the first measured biometric matches a first stored biometric reading. A plurality of first values are gathered prior to the first value. The second biometric reader delivers a second measured biometric for processing by the second biometric matching engine to produce a second value, which is indicative of a likelihood that the second measured biometric matches a second stored biometric reading. A plurality of second values are gathered prior to the second value. The first and second biometric readers measure a different biometric, or the first and second biometric matching engines use a different algorithm. The processor normalizes the first value according to the plurality of first values, normalizes the second value according to the plurality of second values, and determines if the user matches a person using the normalized first and second values.

Claims

1. A biometric system that uses readings from a plurality of biometrics of a user, comprising:

a first biometric reader delivering a first measured biometric;

a first biometric matching engine, wherein;

the first biometric matching engine processes the first measured biometric to produce a first value,

the first value is indicative of a likelihood that the first measured biometric matches a first stored biometric reading, and

a plurality of first values are gathered prior to the first value;

a second biometric reader delivering a second measured biometric;

a second biometric matching engine, wherein;

the second biometric matching engine processes the second measured biometric to produce a second value,

the second value is indicative of a likelihood that the second measured biometric matches a second stored biometric reading,

a plurality of second values are gathered prior to the second value, and

at least one of;

the first and second biometric readers measure a different biometric, or

the first and second biometric matching engines use a different algorithm; and

a processor that;

normalizes the first value according to the plurality of first values,

normalizes the second value according to the plurality of second values, and

determines if the user is matches to a person using the normalized first and second values.

2. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 1, a first threshold is chosen based, at least in part, on the plurality of first values gathered previously by the first biometric reader and the first biometric matching engine.

3. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 2, wherein the first threshold is normalized.

4. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 3, wherein the second threshold is normalized.

5. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 3, wherein the first and second threshold are combined to find a combined threshold.

6. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 2, a second threshold is chosen based, at least in part, on the plurality of second values gathered previously by the second biometric reader and the second biometric matching engine.

7. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 1, wherein a first biometric matching engine algorithm updates as the first value is added to the plurality of first values.

8. The biometric system that uses readings from the plurality of biometrics of the user as recited in claim 1, wherein the first biometric reader retrieves from storage the first measured biometric.

9. A method for increasing accuracy of a biometric matching over time, the method comprising:

gathering a first plurality of biometrics;

matching a first measured biometric with the first plurality of biometrics to produce a first result, wherein;

the first result is indicative of a likelihood that the first measured biometric matches at least one of the first plurality of biometrics, and

the first measured biometric is a different type of biometric than the second measured biometric;

gathering a second plurality of biometrics;

matching a second measured biometric with the second plurality of biometrics to produce a second result, wherein;

the first-listed matching step uses a different algorithm from the second-listed matching step, and

the second result is indicative of a likelihood that the second measured biometric matches at least one of the second plurality of biometrics;

combining the first and second results in a normalizing process;

matching a user to a person using a result from the normalizing process; and

adding the first result to a plurality of first results.

10. The method for increasing accuracy of the biometric authentication over time as recited in claim 9, wherein the first and second measured biometrics are gathered from the user.

11. The method for increasing accuracy of the biometric authentication over time as recited in claim 9, wherein the first plurality of biometrics is gathered from a plurality of biometric readers.

12. The method for increasing accuracy of the biometric authentication over time as recited in claim 9, further comprising a step of adding the second measured biometric to the second plurality of biometrics, whereby a biometric match algorithm of the second-listed matching step improves over time.

13. The method for increasing accuracy of the biometric authentication over time as recited in claim 9, wherein the gathering step comprises a step of latently gathering the first plurality of biometrics, which were previously captured and stored.

14. The method for increasing accuracy of the biometric authentication over time as recited in claim 9, further comprising a step of adding the second value to the plurality of second values after the authenticating step.

15. The method for increasing accuracy of the biometric authentication over time as recited in claim 9, further comprising a step of recalibrating performance of the first-listed matching step with the first measured biometric.

16. A biometric system that uses readings from a plurality of biometrics of a user to authenticate the user, comprising:

a first biometric reader producing a first measured biometric from the user;

a first biometric matching engine, wherein;

the first biometric matching engine processes the first measured biometric to produce a first value, and

the first value is indicative of a likelihood that the first measured biometric matches a first stored biometric reading;

a second biometric reader producing a second measured biometric from the user;

a second biometric matching engine, wherein;

the second biometric matching engine processes the second measured biometric to produce a second value,

the second value is indicative of a likelihood that the second measured biometric matches a second stored biometric reading, and

at least one of;

the first and second biometric readers measure a different biometric, or

the first and second biometric matching engines use a different algorithm; and

a processor that;

adjusts the first and second values to use a common scale, and

determines if the user is authentic using the adjusted first and second values that use the common scale,

wherein a first threshold is chosen based, at least in part, on a plurality of first values gathered with the first biometric reader and the fist biometric matching engine.

17. A biometric system that uses readings from a plurality of biometrics of a user to authenticate the user, comprising:

a first biometric reader producing a first measured biometric from the user;

a first biometric matching engine, wherein;

the first biometric matching engine processes the first measured biometric to produce a first value, and

the first value is indicative of a likelihood that the first measured biometric matches a first stored biometric reading;

a second biometric reader producing a second measured biometric from the user;

a second biometric matching engine, wherein;

the second biometric matching engine processes the second measured biometric to produce a second value,

the second value is indicative of a likelihood that the second measured biometric matches a second stored biometric reading, and

at least one of;

the first and second biometric readers measure a different biometric, or

the first and second biometric matching engines use a different algorithm; and

a processor that;

adjusts the first and second values to use a common scale, and

determines if the user is authentic using the adjusted first and second values that use the common scale,

wherein the biometric system recalibrates as new first and second values are gathered over time.

18. A biometric system that uses readings from a plurality of biometrics of a user to authenticate the user, comprising:

a first biometric reader producing a first measured biometric from the user;

a first biometric matching engine, wherein;

the first biometric matching engine processes the first measured biometric to produce a first value, and

the first value is indicative of a likelihood that the first measured biometric matches a first stored biometric reading;

a second biometric reader producing a second measured biometric from the user;

a second biometric matching engine, wherein;

the second biometric matching engine processes the second measured biometric to produce a second value,

the second value is indicative of a likelihood that the second measured biometric matches a second stored biometric reading, and

at least one of;

the first and second biometric readers measure a different biometric, or

the first and second biometric matching engines use a different algorithm; and

a processor that;

adjusts the first and second values to use a common scale, and

determines if the user is authentic using the adjusted first and second values that use the common scale,

wherein the first biometric reader producing the first measured biometric by retrieving the first measured biometric from a data store where the first measured biometric was previously stored.