Dynamic thresholding for a fingerprint matching system

• US 7,257,241 B2
• Filed: 01/07/2005
• Issued: 08/14/2007
• Est. Priority Date: 01/07/2005
• Status: Expired due to Fees

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Lo, Peter Z.
• Primary Examiner(s): Mehta; Bhavesh M
• Assistant Examiner(s): Kuhn; Jordan
• Application Number:
  US11/031,835
• Publication Number:
  US 20060153433A1
• Time in Patent Office: 949 Days
• Field of Search: 382/124, 382/125
• US Class Current: 382/124
• CPC Class Codes: G06V 40/1365 Matching; Classification
• CPC Class Codes: G11B 20/00086 Circuits for prevention of unauthorised reproduction or copying, e.g. piracy indicating unauthorised use of record carriers in general G11B23/28; scrambling for television signal recording H04N5/913; network architectures or network protocols for network security H04L63/00; cryptographic mechanisms or cryptographic arrangements for secret or secure communication H04L9/00

Abstract

A method for use in print matching apparatus that includes the steps of: enrolling (200) a print of a user; extracting (210) a set of matching features from the enrolled print; comparing (220) the extracted features of the enrolled print to extracted features of each print in a set of reference prints, wherein each reference print in the set is from a person other than the user, and determining a highest non-matching score; determining (230) at least two minimum and maximum threshold pairs based on characteristics of the print matching apparatus; and storing (240) the matching features of the enrolled print, the highest non-matching score, and the at least two maximum and minimum threshold pairs for use in verifying a subsequently captured print of the user.

Claims

1. A method for use in print matching apparatus comprising the steps of:

enrolling a print of a user;

extracting a set of matching features from the enrolled print;

comparing the extracted features of the enrolled print to extracted features of each print in a set of reference prints, wherein each reference print in the set is from a person other than the user, and determining a highest non-matching score;

determining at least two minimum and maximum matched score threshold pairs based on characteristics of the print matching apparatus; and

storing the matching features of the enrolled print, the highest non-matching score, and the at least two maximum and minimum matched score threshold pairs for use in verifying a subsequently captured print of the user, wherein the highest non-matching score and a selected one of the minimum and maximum matched score threshold pairs are used in determining a verification threshold used in verifying the subsequently captured print of the user.

2. The method of claim 1, wherein the step of enrolling includes capturing at least one print of the user and comparing the quality of the at least one captured print to a pre-determined quality threshold for selecting one of the at least one captured prints.

3. The method of claim 1, wherein each set of matching features includes a set of minutiae.

4. The method of claim 1, wherein the set of reference prints has a first number of reference prints that is selected based on an accuracy requirement, and the highest non-matching score is based on the first number of prints.

5. The method of claim 4, wherein the highest non-matching score is based on a second number of reference prints and is generated by one of interpolating and extrapolating the results of comparing the extracted features of the enrolled print to extracted features of each print in the first number of reference prints.

6. The method of claim 1, wherein each minimum and maximum matched score threshold pair is determined from a different characteristic curve of the print matcher apparatus.

7. The method of claim 6, wherein each characteristic curve is a Receiver Operator Characteristic (ROC) curve that is generated based on a quality measure and a classification type.

8. The method of claim 7, wherein the classification type is selected from a group comprising a whorl, a left loop, a right loop, an arch and an unknown pattern.

9. The method of claim 7, wherein each minimum matched score threshold is a score point on the corresponding ROC curve that satisfies a maximum false hit accuracy requirement, and each maximum match score threshold is a score point on the corresponding ROC curve that satisfies a minimum false hit accuracy requirement.

19. Print matching apparatus configured for performing the method of claim 1.

10. A method for use in print matching apparatus comprising the steps of:

capturing a print;

extracting a set of matching features from the captured print;

retrieving matching features of a previously enrolled print of a user and a highest non-matching score for the user;

selecting one minimum and maximum matched score threshold pair from a set of at least two pre-determined minimum and maximum matched score threshold pairs, wherein each minimum and maximum matched score threshold pair is based on characteristics of the print matcher apparatus;

determining a verification threshold based at least on the highest non-matching score and the selected one minimum and maximum matched score threshold pair; and

comparing the matching features of the captured print to the retrieved matching features based on the verification threshold to determine whether the captured print is that of the user.

11. The method of claim 10, wherein each set of matching features includes a set of minutiae having a corresponding number of minutiae, the method further comprising the step of determining a bias score based on the difference between the number of minutiae in the set of matching features of the captured print and the number of minutiae in the set of retrieved matching features.

12. The method of claim 11, wherein the verification threshold is determined further based on the bias score.

13. The method of claim 12, wherein the verification threshold is set to a function of the highest non-matching score plus the bias score, bounded by the selected maximum and minimum matched score threshold pair.

14. The method of claim 11, wherein:

a normalized score difference is determined based on the difference in the number of minutiae between the matching features of the captured print and the retrieved matching features;

a normalized score difference is determined for a plurality of captured prints whose corresponding matching features are compared to the retrieved matching features; and

the bias score is determined as a function of at least one of the mean and standard deviation of all of the determined normalized score differences.

15. The method of claim 10, wherein the minimum and maximum matched score threshold pair is selected based on a quality measure and a classification determination for the captured print and the previously enrolled print.

16. The method of claim 15, wherein the classification determination is selected from a group comprising a whorl, a left loop, a right loop, an arch and an unknown pattern.

17. The method of claim 10, wherein the step of capturing includes capturing at least one print and comparing the quality of the at least one captured print to a pre-determined quality threshold for selecting one of the at least one captured prints.

18. The method of claim 10, wherein:

a matched score is generated by comparing the matching features of the captured print to the retrieved matching features;

a function of the matched score is compared to the verification threshold; and

a match is determined when the function of the matched score is at least equal to the verification threshold.

20. Print matching apparatus configured for performing the method of claim 10.