Multi-view fingerprint matching

• US 10,013,597 B2
• Filed: 11/14/2016
• Issued: 07/03/2018
• Est. Priority Date: 08/11/2014
• Status: Active Grant

• Current Assignee: Synaptics Incorporated
• Original Assignee: Synaptics Incorporated
• Inventors: Russo, Anthony P., Krishnapura, Rohini
• Primary Examiner(s): WERNER, BRIAN P
• Application Number:
  US15/351,299
• Publication Number:
  US 20170061196A1
• Time in Patent Office: 596 Days
• CPC Class Codes: G06V 40/1335 Combining adjacent partial images (e.g. slices) to create a composite input or reference pattern; Tracking a sweeping finger movement
• CPC Class Codes: G06V 40/1365 Matching; Classification
• CPC Class Codes: G06V 40/1371 Matching features related to minutiae or pores

Abstract

A method and a device are provided for performing a recognition process. The recognition process compares an individual fingerprint view to a fingerprint enrollment template in order to determine whether a match has been found. The determination of a match is based on individual match statistics collected between the individual fingerprint view and each view of the fingerprint enrollment template. Additionally, inter-view match statistics between each view of the fingerprint enrollment template may also be determined. The inter-view match statistics can be analyzed along with the individual match statistics to further inform the determination of a match between the individual fingerprint view and the fingerprint enrollment template.

Claims

1. A method of biometric matching to an enrollment template, the method comprising:

comparing a verification view of a verification template to a plurality of individual enrollment views of the enrollment template to determine individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views, wherein the verification view corresponds to a biometric image captured by an input device, wherein the plurality of individual enrollment views correspond to a plurality of different biometric images captured by the input device;

determining at least one inter-view transformation between at least one pairing of the individual enrollment views, wherein the at least one inter-view transformation includes at least one translation and at least one rotation;

determining feature vectors from the individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views;

calculating a composite match score between the verification template and the enrollment template by inputting the feature vectors into a machine learning classifier; and

indicating a biometric match between the verification template and the enrollment template based on the composite match score.

2. The method of claim 1, wherein at least some of the feature vectors are determined from inter-view match statistics between at least one pairing of the individual enrollment views within the enrollment template.

3. The method of claim 1, wherein at least some of the feature vectors are determined from at least one inter-view geometric transformation between at least one pairing of the individual enrollment views within the enrollment template.

4. The method of claim 1, wherein the feature vectors include a transformation error.

6. The method of claim 1, wherein the individual match statistics include a number of matched points.

7. The method of claim 1, wherein the individual match statistics include a number of non-matched points.

8. The method of claim 1, wherein the machine learning classifier comprises a neural network.

9. The method of claim 2, wherein the biometric match is indicated in response to determining that the composite match score satisfies a threshold.

10. The method of claim 1, further comprising:

capturing the biometric image with the input device.

11. The method of claim 1, wherein the plurality of individual enrollment views are all individual enrollment views contained in the enrollment template.

12. The method of claim 2, wherein the plurality of individual enrollment views are a subset of all individual enrollment views contained in the enrollment template.

26. The method of claim 1, wherein the input device is only large enough to capture a partial image of a biometric sample, and wherein the plurality of different biometric images correspond to a plurality of different partial images captured by the input device.

5. A method of biometric matching to an enrollment template, the method comprising:

comparing a verification view of a verification template to a plurality of individual enrollment views of the enrollment template to determine individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views, wherein the verification view corresponds to a biometric sample captured by an input device;

determining feature vectors from the individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views;

calculating a composite match score between the verification template and the enrollment template by inputting the feature vectors into a machine learning classifier; and

indicating a biometric match between the verification template and the enrollment template based on the composite match score,

wherein the feature vectors include a transformation error, and

wherein the method further comprises computing the transformation error from at least one inter-view geometric transformation between at least one pairing of the individual enrollment views within the enrollment template.

13. A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, causes a computing device to perform a method comprising:

comparing a verification view of a verification template to a plurality of individual enrollment views of the enrollment template to determine individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views, wherein the verification view corresponds to a biometric image captured by an input device, wherein the plurality of individual enrollment views correspond to a plurality of different biometric images captured by the input device;

determining at least one inter-view transformation between at least one pairing of the individual enrollment views, wherein the at least one inter-view transformation includes at least one translation and at least one rotation;

determining feature vectors from the individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views;

calculating a composite match score between the verification template and the enrollment template by inputting the feature vectors into a machine learning classifier; and

indicating a biometric match between the verification template and the enrollment template based on the composite match score.

14. The non-transitory computer-readable storage medium of claim 13, wherein at least some of the feature vectors are determined from inter-view match statistics between at least one pairing of the individual enrollment views within the enrollment template.

15. The non-transitory computer-readable storage medium of claim 13, wherein at least some of the feature vectors are determined from at least one inter-view geometric transformation between at least one pairing of the individual enrollment views within the enrollment template.

16. The non-transitory computer-readable storage medium of claim 13, wherein the feature vectors include a transformation error.

18. The non-transitory computer-readable storage medium of claim 13, wherein the individual match statistics include a number of matched points.

19. The non-transitory computer-readable storage medium of claim 13, wherein the individual match statistics include a number of non-matched points.

20. The non-transitory computer-readable storage medium of claim 13, wherein the machine learning classifier comprises a neural network.

21. The non-transitory computer-readable storage medium of claim 14, wherein the biometric match is indicated in response to determining that the composite match score satisfies a threshold.

27. The non-transitory computer-readable storage medium of claim 13, wherein the input device is only large enough to capture a partial image of a biometric sample, and wherein the plurality of different biometric images correspond to a plurality of different partial images captured by the input device.

17. A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, causes a computing device to perform a method comprising:

comparing a verification view of a verification template to a plurality of individual enrollment views of the enrollment template to determine individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views, wherein the verification view corresponds to a biometric sample captured by an input device;

determining feature vectors from the individual match statistics between the verification view and each individual enrollment view of the plurality of individual enrollment views;

calculating a composite match score between the verification template and the enrollment template by inputting the feature vectors into a machine learning classifier; and

indicating a biometric match between the verification template and the enrollment template based on the composite match score,

wherein the feature vectors include a transformation error, and

wherein the method further comprises computing the transformation error from at least one inter-view geometric transformation between at least one pairing of the individual enrollment views within the enrollment template.

22. A device for fingerprint matching to an enrollment template, the device comprising:

a fingerprint sensor; and

a processor configured to;

compare a fingerprint view of a fingerprint template to a plurality of individual enrollment views of the enrollment template to determine individual match statistics between the fingerprint view and each individual enrollment view of the plurality of individual enrollment views, wherein the fingerprint view corresponds to a fingerprint image captured by the fingerprint sensor, wherein the plurality of individual enrollment views correspond to a plurality of different fingerprint images captured by the fingerprint sensor;

determine at least one inter-view transformation between at least one pairing of the individual enrollment views, wherein the at least one inter-view transformation includes at least one translation and at least one rotation;

determine feature vectors from the individual match statistics between the fingerprint view and each individual enrollment view of the plurality of individual enrollment views;

calculate a composite match score between the fingerprint template and the enrollment template by inputting the feature vectors into a machine learning classifier; and

indicate a fingerprint match between the fingerprint template and the enrollment template based on the composite match score.

23. The device of claim 22, wherein the feature vectors include a transformation error.

24. The device of claim 22, wherein the individual match statistics include a number of matched points.

25. The device of claim 22, wherein the individual match statistics include a number of non-matched points.

28. The device of claim 22, wherein the fingerprint sensor is only large enough to capture a partial image of a fingerprint, and wherein the plurality of different fingerprint images correspond to a plurality of different partial images captured by the fingerprint sensor.