Systems and methods for spoof detection based on local binary patterns

1. A device, comprising:

a biometric sensor; and

a processing system configured to:

receive, from the biometric sensor, an input image of a biometric;

analyze the input image to identify pixels associated with a first feature in the input image, wherein the input image comprises an image of a fingerprint, and wherein the pixels associated with the first feature comprise fingerprint ridge contour pixels at border pixels where a fingerprint ridge meets a fingerprint valley in the fingerprint;

perform a local binary pattern (LBP) analysis on the fingerprint ridge contour pixels, comprising computing, for a given pixel of the fingerprint ridge contour pixels, an LBP value corresponding to texture of one or more neighboring pixels around the given pixel; and

determine whether the input image is a replica of the biometric based on results of performing the LBP analysis on the fingerprint ridge contour pixels, wherein the LBP analysis is performed on less than all pixels in the input image.

2. The device of claim 1 , wherein the processing system is further configured to:

identify ridge pixels in the image of the fingerprint, wherein the pixels associated with the first feature in the input image further comprise the ridge pixels in the image of the fingerprint, and wherein determining whether the input image is a replica of the biometric is further based on results of performing the LBP analysis on the ridge pixels.

3. The device of claim 1 , wherein the processing system is further configured to:

analyze the input image to identify pixels associated with a second feature in the input image; and

perform an LBP analysis on the pixels associated with the second feature;

wherein determining whether the input image is a replica of the biometric is further based on results of performing the LBP analysis on the pixels associated with the second feature.

4. The device of claim 3 , wherein pixels associated with the second feature comprise one or more of:

pixels associated with fingerprint ridge skeletons;

pixels associated with fingerprint valleys;

pixels associated with fingerprint minutiae; and

pixels associated with fingerprint keypoints.

5. The device of claim 3 , wherein the LBP analysis performed on the fingerprint ridge contour pixels is the same as the LBP analysis performed on the pixels associated with the second feature.

6. The device of claim 3 , wherein the LBP analysis performed on the fingerprint ridge contour pixels is different than the LBP analysis performed on the pixels associated with the second feature.

7. The device of claim 1 , wherein the processing system is further configured to:

generate a statistical distribution of the LBP values of the fingerprint ridge contour pixels;

wherein determining whether the input image is a replica of the biometric is based on the statistical distribution.

8. The device of claim 7 , wherein the statistical distribution comprises a histogram.

9. The device of claim 1 , wherein the processing system is further configured to:

perform an LBP analysis on one or more enrollment images of the biometric stored in a template storage;

wherein determining whether the input image is a replica of the biometric is further based on a differential between results of performing the LBP analysis on the fingerprint ridge contour pixels and results of performing the LBP analysis on the one or more enrollment images of the biometric.

10. The device of claim 1 , wherein performing the LBP analysis on the fingerprint ridge contour pixels is performed after determining that the input image comprises a match to one or more enrollment images of the biometric stored in a template storage.

11. The device of claim 1 , wherein the replica comprises a gelatin mold, a graphite mold, or a wood glue mold of the fingerprint of a finger.

12. The device of claim 1 , wherein the neighboring pixels around the given pixel comprise eight pixels around the given pixel, including pixels on a top, a bottom, a left, a right, and at four corners of a location of the given pixel.

13. The device of claim 1 , wherein the pixels associated with the first feature further comprise pixels associated medial axes of fingerprint ridges of the fingerprint, and wherein determining whether the input image is a replica of the biometric is further based on results of performing the LBP analysis on the pixels associated medial axes of fingerprint ridges of the fingerprint.

14. A method for performing spoof detection, comprising:

receiving, by a processor from a biometric sensor, an input image of a biometric;

analyzing, by the processor, the input image to identify pixels associated with a first feature in the input image, wherein the input image comprises an image of a fingerprint, and wherein the pixels associated with the first feature comprise fingerprint ridge contour pixels at border pixels where a fingerprint ridge meets a fingerprint valley in the fingerprint;

performing, by the processor, a local binary pattern (LBP) analysis on the fingerprint ridge contour pixels, which comprises computing, for a given pixel of the fingerprint ridge contour pixels, an LBP value corresponding to texture of one or more neighboring pixels around the given pixel, wherein the neighboring pixels around the given pixel comprise eight pixels around the given pixel, including pixels on a top, a bottom, a left, a right, and at four corners of a location of the given pixel; and

determining, by the processor, whether the input image is a replica of the biometric based on results of performing the LBP analysis on the fingerprint ridge contour pixels, wherein the LBP analysis is performed on less than all pixels in the input image.

15. The method of claim 14 , further comprising:

identifying ridge pixels in the fingerprint, wherein determining whether the input image is a replica of the biometric is further based on results of performing the LBP analysis on the ridge pixels.

16. The method of claim 14 , further comprising:

analyzing the input image to identify pixels associated with a second feature in the input image; and

performing an LBP analysis on the pixels associated with the second feature;

wherein determining whether the input image is a replica of the biometric is further based on results of performing the LBP analysis on the pixels associated with the second feature.

17. The method of claim 16 , wherein pixels associated with the second feature comprise one or more of:

pixels associated with fingerprint ridge skeletons;

pixels associated with fingerprint valleys;

pixels associated with fingerprint minutiae; and

pixels associated with fingerprint keypoints.

18. The method of claim 16 , wherein the LBP analysis performed on the fingerprint ridge contour pixels is the same as the LBP analysis performed on the pixels associated with the second feature, or the LBP analysis performed on the fingerprint ridge contour pixels is different than the LBP analysis performed on the pixels associated with the second feature.

19. The method of claim 14 , wherein performing the LBP analysis comprises performing rotation-invariant LBP analysis or performing the LBP analysis comprises performing rotation-dependent LBP analysis.

20. The method of claim 14 , wherein performing the LBP analysis on the fingerprint ridge contour pixels is performed after determining that the input image comprises a match to one or more enrollment images of the biometric stored in a template storage, the method further comprising:

performing an LBP analysis on one or more enrollment images of the biometric stored in a template storage, wherein determining whether the input image is a replica of the biometric is further based on a differential between results of performing the LBP analysis on the fingerprint ridge contour pixels and results of performing the LBP analysis on the one or more enrollment images of the biometric.

21. A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, causes a computing device to perform spoof detection, by performing steps comprising:

receiving, from a biometric sensor, an input image of a biometric;

analyzing the input image to identify pixels associated with a first feature in the input image, wherein the input image comprises an image of a fingerprint, and wherein the pixels associated with the first feature comprise fingerprint ridge contour pixels at border pixels where a fingerprint ridge meets a fingerprint valley in the fingerprint;

performing a local binary pattern (LBP) analysis on the fingerprint ridge contour pixels, which comprises computing, for a given pixel of the fingerprint ridge contour pixels, an LBP value corresponding to texture of one or more neighboring pixels around the given pixel; and

determining whether the input image is a replica of the biometric based on results of performing the LBP analysis on the fingerprint ridge contour pixels, wherein the LBP analysis is performed on less than all pixels in the input image.