Minutiae grouping for distorted fingerprint matching

1. A method for matching distorted fingerprints, the method implemented by an automatic fingerprint identification system including a processor, a memory coupled to the processor, an interface to a fingerprint scanning device, and a sensor associated with the fingerprint scanning device that indicates a fingerprint match, the method comprising:

selecting (i) a list of search minutia extracted from a search fingerprint, and (ii) a list of reference minutia extracted from a reference fingerprint;

for each particular search minutia included in the list of search minutiae, identifying a corresponding reference minutia from the list of reference minutiae that is a closest matched minutia to the particular search minutia;

generating a first list of locally matched minutiae between the search fingerprint and the reference fingerprint that includes each of the particular search minutiae, and each corresponding reference minutia that is the closest matched minutia to each of the particular search minutiae;

identifying a first set of globally aligned minutiae pairs from among the first list of locally matched minutiae pairs based at least on performing a global alignment procedure on a first subset of the first list of locally matched minutiae pairs;

computing (i) a first similarity score between the list of particular search minutiae and the list of corresponding reference minutiae within the first set of globally aligned minutiae pairs, and (ii) a center coordinate for the first set of globally aligned minutiae pairs;

generating a second list of locally matched minutiae pairs, based at least on removing each of the particular search minutiae and each corresponding reference minutia that is the closest matched minutia to each of the particular search minutiae, that are included in the first set of globally aligned minutiae pairs, from the first list of locally matched minutiae pairs;

identifying a second set of globally aligned minutiae pairs from among the second list of locally matched minutiae pairs based at least on performing the global alignment procedure on a second subset of the second list of locally matched minutiae pairs;

determining that a number of globally aligned minutiae pairs within the second set of globally aligned minutiae pairs exceeds a threshold value;

computing (i) a second similarity score between the list of particular search minutiae and the list of corresponding reference minutiae within the second set of globally aligned minutiae pairs, and (ii) a center coordinate for the second set of globally aligned minutiae pairs;

identifying one or more additional sets of globally aligned minutiae pairs from among the second list of locally matched minutiae pairs based at least on successively performing the global alignment procedure on one or more additional subsets of the second list of locally matched minutiae pairs until the second list of locally matched minutiae pairs does not contain any particular search minutia, wherein:

each successive individual subset from the one or more additional subsets of the second list of locally matched minutia pairs includes particular search minutiae and the corresponding reference minutiae that are not included in a prior individual subset that was previously identified based at least on performing a prior global alignment procedure, and

after performing the global alignment procedure on each successive individual subset, removing each of the particular search minutiae and each of the corresponding reference minutiae that are included in the prior individual subset from the second list of locally matched minutiae pairs;

computing (i) additional similarity scores between the list of particular search minutiae and the list of corresponding reference minutiae within each of the one or more additional sets of globally aligned minutiae pairs of globally aligned minutiae pairs, and (ii) a center coordinate for each of the one or more additional sets of globally aligned minutiae pairs of globally aligned minutiae pairs;

determining a geometric consistency metric between (i) the center coordinate for the first set of globally aligned minutiae pairs, (ii) the center coordinate for the second set of globally aligned minutiae pairs, and (iii) the center coordinates for each of the one or more additional sets of globally aligned minutiae pairs;

computing a final similarity score based at least on (i) combining the first similarity score, the second similarity score, and the additional similarity scores, and (ii) the geometric consistency metric; and

providing, for output to the automatic fingerprint identification system, the final similarity score for use in a particular fingerprint matching operation between the search fingerprint and the reference fingerprint.

2. The method of claim 1 , wherein generating a first list of locally matched minutiae pairs between the search fingerprint and the reference fingerprint comprises:

constructing, for each particular search minutia, a local geometric structure; and

identifying, for each particular search minutia, a corresponding reference minutiae that is a closest matched minutia based at least on the local geometric structure for each particular search minutiae.

3. The method of claim 1 , wherein identifying a first set of globally aligned minutiae pairs comprises:

computing (i) a set of rotation parameters, and (ii) a set of translation parameters, based at least on closest locally matched minutiae pairs within the first list of locally matched minutiae; and

aligning each particular search minutiae within the first subset of the first list of locally matched minutiae pairs to the corresponding reference minutia based at least on the set of rotation parameters and the set of translation parameters.

4. The method of claim 3 , comprising:

identifying a set of best aligned minutiae pairs from among the first set of globally aligned minutiae pairs using the set of rotation parameters and the set of translation parameters for each of the search fingerprint and the reference fingerprint.

5. The method of claim 1 , wherein each of the first set of globally aligned minutiae pairs and the second set of globally aligned minutiae pairs include at least three matched minutiae pairs.

6. The method of claim 1 , wherein computing the final similarity score comprises:

determining that the center coordinate of a particular set of globally aligned minutiae pairs from among the one or more additional sets of globally aligned minutiae pairs is geometrically consistent with the center coordinates of each of the first set of globally aligned minutiae pairs;

determining that the second set of globally aligned minutiae pairs of search fingerprint respective to the center coordinate of the particular set of globally aligned minutiae pairs is geometrically consistent with the center coordinates of each of the first set of globally aligned minutiae pairs and the second set of globally aligned minutiae pairs of the reference fingerprint and

computing a revised similarity score based at least on combining the first similarity score, the second similarity score, and a set of revised additional similarity scores that does not include the similarity score of the particular set of globally aligned minutiae pairs.

7. The method of claim 6 , wherein the geometric consistency determination is based at least on one of:

comparing a distance difference between any two centers of the search fingerprint and any two centers of reference fingerprint to a pre-defined parameter for the distance difference, or

comparing an angle difference between any three centers of the search fingerprint and any three centers of reference fingerprint to a pre-defined parameter for the angle.

8. The method of claim 7 , comprising:

determining that the distance difference for a list of search minutiae within a particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the distance difference; and

excluding the list of search minutiae from the particular set of globally aligned minutiae pairs in response to determining that the distance difference for the list of search minutiae within the particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the distance difference.

9. The method of claim 7 , comprising:

determining that the angle difference for a list of search minutiae within a particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the angle difference; and

excluding the list of search minutiae from the particular set of globally aligned minutiae pairs in response to determining that the angle difference for the list of search minutiae within the particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the angle difference.

10. The method of claim 7 , comprising calculating a match similarity score based at least on the geometric consistency determination.

11. A system comprising:

a fingerprint scanning device, including (i) an interface to a fingerprint scanner, and (ii) a sensor associated with the fingerprint scanner, configured to identify a match between a search record representing a fingerprint of a subject and a reference record from among a plurality of reference records representing fingerprints of a plurality of historical subjects;

a database containing the plurality of reference records representing fingerprints of the plurality of historical subjects; and

a computing device in communication with said database and said fingerprint scanning device, said computing device having a processor and a memory coupled to said processor, said computing device configured perform operations comprising:

selecting (i) a list of search minutia extracted from a search fingerprint, and (ii) a list of reference minutia extracted from a reference fingerprint;

for each particular search minutia included in the list of search minutiae, identifying a corresponding reference minutia from the list of reference minutiae that is a closest matched minutia to the particular search minutia;

generating a first list of locally matched minutiae between the search fingerprint and the reference fingerprint that includes each of the particular search minutiae, and each corresponding reference minutia that is the closest matched minutia to each of the particular search minutiae;

identifying a first set of globally aligned minutiae pairs from among the first list of locally matched minutiae pairs based at least on performing a global alignment procedure on a first subset of the first list of locally matched minutiae pairs;

computing (i) a first similarity score between the list of particular search minutiae and the list of corresponding reference minutiae within the first set of globally aligned minutiae pairs, and (ii) a center coordinate for the first set of globally aligned minutiae pairs;

generating a second list of locally matched minutiae pairs, based at least on removing each of the particular search minutiae and each corresponding reference minutia that is the closest matched minutia to each of the particular search minutiae, that are included in the first set of globally aligned minutiae pairs, from the first list of locally matched minutiae pairs;

identifying a second set of globally aligned minutiae pairs from among the second list of locally matched minutiae pairs based at least on performing the global alignment procedure on a second subset of the second list of locally matched minutiae pairs;

determining that a number of globally aligned minutiae pairs within the second set of globally aligned minutiae pairs exceeds a threshold value;

computing (i) a second similarity score between the list of particular search minutiae and the list of corresponding reference minutiae within the second set of globally aligned minutiae pairs, and (ii) a center coordinate for the second set of globally aligned minutiae pairs;

identifying one or more additional sets of globally aligned minutiae pairs from among the second list of locally matched minutiae pairs based at least on successively performing the global alignment procedure on one or more additional subsets of the second list of locally matched minutiae pairs until the second list of locally matched minutiae pairs does not contain any particular search minutia, wherein:

each successive individual subset from the one or more additional subsets of the second list of locally matched minutia pairs includes particular search minutiae and the corresponding reference minutiae that are not included in a prior individual subset that was previously identified based at least on performing a prior global alignment procedure, and

after performing the global alignment procedure on each successive individual subset, removing each of the particular search minutiae and each of the corresponding reference minutiae that are included in the prior individual subset from the second list of locally matched minutiae pairs;

computing (i) additional similarity scores between the list of particular search minutiae and the list of corresponding reference minutiae within each of the one or more additional sets of globally aligned minutiae pairs of globally aligned minutiae pairs, and (ii) a center coordinate for each of the one or more additional sets of globally aligned minutiae pairs of globally aligned minutiae pairs;

determining a geometric consistency metric between (i) the center coordinate for the first set of globally aligned minutiae pairs, (ii) the center coordinate for the second set of globally aligned minutiae pairs, and (iii) the center coordinates for each of the one or more additional sets of globally aligned minutiae pairs;

computing a final similarity score based at least on (i) combining the first similarity score, the second similarity score, and the additional similarity scores, and (ii) the geometric consistency metric; and

providing, for output to the automatic fingerprint identification system, the final similarity score for use in a particular fingerprint matching operation between the search fingerprint and the reference fingerprint.

12. The system of claim 11 , wherein generating a first list of locally matched minutiae pairs between the search fingerprint and the reference fingerprint comprises:

constructing, for each particular search minutia, a local geometric structure; and

identifying, for each particular search minutia, a corresponding reference minutiae that is a closest matched minutia based at least on the local geometric structure for each particular search minutiae.

13. The system of claim 11 , wherein identifying a globally aligned minutiae pairs comprises:

computing (i) a set of rotation parameters, and (ii) a set of translation parameters, based at least on closest locally matched minutiae pairs within the first list of locally matched minutiae; and

aligning each particular search minutiae within the first subset of the first list of locally matched minutiae pairs to the corresponding reference minutia based at least on the set of rotation parameters and the set of translation parameters.

14. The system of claim 13 , wherein the operations comprise:

identifying a set of best aligned minutiae pairs from among the first set of globally aligned minutiae pairs using the set of rotation parameters and the set of translation parameters for each of the search fingerprint and the reference fingerprint.

15. The system of claim 11 , wherein each of the first set of globally aligned minutiae pairs and the second set of globally aligned minutiae pairs include at least three matched minutiae pairs.

16. The system of claim 11 , wherein computing the final similarity score comprises:

determining that the center coordinate of a particular set of globally aligned minutiae pairs from among the one or more additional sets of globally aligned minutiae pairs is geometrically consistent with the center coordinates of each of the first set of globally aligned minutiae pairs;

determining that the second set of globally aligned minutiae pairs of search fingerprint respective to the center coordinate of the particular set of globally aligned minutiae pairs is geometrically consistent with the center coordinates of each of the first set of globally aligned minutiae pairs and the second set of globally aligned minutiae pairs of the reference fingerprint and

computing a revised similarity score based at least on combining the first similarity score, the second similarity score, and a set of revised additional similarity scores that does not include the similarity score of the particular set of globally aligned minutiae pairs.

17. The system of claim 16 , wherein the geometric consistency determination is based at least on one of:

comparing a distance difference between any two centers of the search fingerprint and any two centers of reference fingerprint to a pre-defined parameter for the distance difference, or

comparing an angle difference between any three centers of the search fingerprint and any three centers of reference fingerprint to a pre-defined parameter for the angle.

18. The system of claim 17 , wherein the operations comprise:

determining that the distance difference for a list of search minutiae within a particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the distance difference; and

excluding the list of search minutiae from the particular set of globally aligned minutiae pairs in response to determining that the distance difference for the list of search minutiae within the particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the distance difference.

19. The system of claim 17 , wherein the operations comprise:

determining that the angle difference for a list of search minutiae within a particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the angle difference; and

excluding the list of search minutiae from the particular set of globally aligned minutiae pairs in response to determining that the angle difference for the list of search minutiae within the particular set of globally aligned minutiae pair does not satisfy the pre-defined parameter for the angle difference.

20. The system of claim 17 , wherein the operations comprise:

calculating a match similarity score based at least on the geometric consistency determination.