Texture features for biometric authentication
This specification describes technologies relating to biometric authentication based on images of the eye. In general, one aspect of the subject matter described in this specification can be embodied in methods that include obtaining one or more image regions from a first image of an eye. Each of the image regions may include a view of a respective portion of the white of the eye. The method may further include applying several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region. The several distinct filters may include convolutional filters that are each configured to describe one or more aspects of an eye vasculature and in combination describe a visible eye vasculature in a feature space. A match score may be determined based on the generated descriptors and based on one or more descriptors associated with a second image of eye vasculature.
1. A computer-implemented method comprising:
obtaining one or more image regions from a first image of an eye, wherein each of the image regions includes a view of a respective portion of vasculature of the eye external to a corneal limbus boundary of the eye;
applying several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region, wherein the several distinct filters comprises a plurality of convolutional filters that are each configured to describe one or more aspects of an eye vasculature and in combination describe a visible eye vasculature in a feature space; and
determining a match score based on the generated descriptors and based on one or more descriptors associated with a second image of eye vasculature.
2. The method of claim 1 , in which the visible eye vasculature is described by some of the generated descriptors that are based on co-occurrence statistics of image data elements in one or more of the image regions.
3. The method of claim 1 , in which the visible eye vasculature is described by some of the generated descriptors that are based on information-theoretic statistics of image data elements in one or more of the image regions.
4. The method of claim 1 , in which one of the convolutional filters is a Gabor filter.
5. The method of claim 1 , in which a filter in the several filters is a non-linear filter.
6. The method of claim 5 , in which the non-linear filter is a trained neural network.
7. The method of claim 1 , in which applying the several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region, further comprises combining respective outputs of one or more of the filters.
8. The method of claim 1 , in which determining a match score further comprises:
combining the respective descriptors for each of the regions into a respective vector for the region;
for one or more of the regions of the first image, comparing the respective vector for the region to a vector derived from descriptors for a corresponding region of the second image to generate a respective similarity score; and
determining the match score based on, at least, the generated similarity scores.
9. The method of claim 1 , further comprising determining whether the match score exceeds a value, wherein the value is based on, at least, a robust threshold in a neighborhood of a three-dimensional receiver operating curve generated from a sensitivity analysis.
10. The method of claim 1 , in which obtaining the one or more image regions comprises tiling a region to obtain smaller regions.
11. A system, comprising:
a data processing apparatus; and
a memory coupled to the data processing apparatus having instructions stored thereon which, when executed by the data processing apparatus cause the data processing apparatus to perform operations comprising:
obtaining one or more image regions from a first image of an eye, wherein each of the image regions includes a view of a respective portion of vasculature of the eye external to a corneal limbus boundary of the eye;
applying several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region, wherein the several distinct filters comprises a plurality of convolutional filters that are each configured to describe one or more aspects of an eye vasculature and in combination describe a visible eye vasculature in a feature space; and
determining a match score based on the generated descriptors and based on one or more descriptors associated with a second image of eye vasculature.
12. The system of claim 11 , in which the visible eye vasculature is described by some of the generated descriptors that are based on co-occurrence statistics of image data elements in one or more of the image regions.
13. The system of claim 11 , in which the visible eye vasculature is described by some of the generated descriptors that are based on information-theoretic statistics of image data elements in one or more of the image regions.
14. The system of claim 11 , in which one of the convolutional filters is a Gabor filter.
15. The system of claim 11 , in which a filter in the several filters is a non-linear filter.
16. The system of claim 15 , in which the non-linear filter is a trained neural network.
17. The system of claim 11 , in which applying the several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region, further comprises combining respective outputs of one or more of the filters.
18. The system of claim 11 , in which determining a match score further comprises:
combining the respective descriptors for each of the regions into a respective vector for the region;
for one or more of the regions of the first image, comparing the respective vector for the region to a vector derived from descriptors for a corresponding region of the second image to generate a respective similarity score; and
determining the match score based on, at least, the generated similarity scores.
19. The system of claim 11 , in which the operations further comprise determining whether the match score exceeds a value, wherein the value is based on, at least, a robust threshold in a neighborhood of a three-dimensional receiver operating curve generated from a sensitivity analysis.
20. The system of claim 11 , in which obtaining the one or more image regions comprises tiling a region to obtain smaller regions.
21. A system, comprising:
a module configured to obtain one or more image regions from a first image of an eye, wherein each of the image regions includes a view of a respective portion of vasculature of the eye external to a corneal limbus boundary of the eye;
a means for applying several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region, wherein the several distinct filters comprises a plurality of convolutional filters that are each configured to describe one or more aspects of an eye vasculature and in combination describe a visible eye vasculature in a feature space; and
a module configured to determine a match score based on the generated descriptors and based on one or more descriptors associated with a second image of eye vasculature.
22. The system of claim 21 , in which the visible eye vasculature is described by some of the generated descriptors that are based on co-occurrence statistics of image data elements in one or more of the image regions.
23. The system of claim 21 , in which the visible eye vasculature is described by some of the generated descriptors that are based on information-theoretic statistics of image data elements in one or more of the image regions.
24. The system of claim 21 , in which one of the convolutional filters is a Gabor filter.
25. The system of claim 21 , in which a filter in the several filters is a non-linear filter.
26. The system of claim 25 , in which the non-linear filter is a trained neural network.
27. The system of claim 21 , in which applying the several distinct filters to each of the image regions to generate a plurality of respective descriptors for the region, further comprises combining respective outputs of one or more of the filters.
28. The system of claim 21 , in which determining a match score further comprises:
combining the respective descriptors for each of the regions into a respective vector for the region;
for one or more of the regions of the first image, comparing the respective vector for the region to a vector derived from descriptors for a corresponding region of the second image to generate a respective similarity score; and
determining the match score based on, at least, the generated similarity scores.
29. The system of claim 21 , further comprising a module configured to determine whether the match score exceeds a value, wherein the value is based on, at least, a robust threshold in a neighborhood of a three-dimensional receiver operating curve generated from a sensitivity analysis.
30. The system of claim 21 , in which obtaining the one or more image regions comprises tiling a region to obtain smaller regions.