Quality metrics for biometric authentication

1. A computer-implemented method comprising:

obtaining a plurality of images of an eye, wherein the images each include a view of a portion of a vasculature of the eye;

determining a plurality of respective metrics for each of the images based on analyzing image data elements of the image, wherein the respective metrics include a first metric that represents a calculated area of one or more connected structures of programmatically enhanced vasculature in the image and a second metric that represents an extent of detected vasculature in the image based on a plurality of color difference signals determined from color components of the image;

calculating a respective first score for each of the images wherein the first score represents a comparison of the respective metrics of the image with corresponding metrics for a reference image;

calculating a second score based on, at least, a combination of the first scores; and

granting a user access to a device or a service based on the second score.

2. The method of claim 1 , further comprising:

before calculating the respective first score for each of the images, determining a respective quality score for each of the images based on the metrics, wherein the quality score is a prediction of the second score assuming the image and the reference image include a view of a same person's vasculature.

3. The method of claim 2 , further comprising:

accepting or rejecting the image based on the respective quality score.

4. The method of claim 1 wherein calculating the second score based on, at least, the combination of the first scores comprises:

determining a respective weight for each of the first scores; and

combining the first scores weighted by their respective weights to determine the second score.

5. The method of claim 1 , wherein calculating a respective first score for each of the images comprises:

a distance between the metrics for the image and the metrics for the reference image.

6. The method of claim 5 wherein the distance is a Euclidean distance, a correlation coefficient, a modified Hausdorff distance, a Mahalanobis distance, a Bregman divergence, a cosine similarity, a Kullback-Leibler distance, or a Jensen-Shannon divergence.

7. The method of claim 1 wherein determining the first metric comprises:

dilating the vasculature in the image so that one or more disconnected veins in the vasculature become connected;

thinning the dilated vasculature in the image; and

determining a portion of the image that contains the thinned dilated vasculature.

8. The method of claim 1 wherein the second metric is for comparing an extent of eye vasculature detected across different color components in the image, and wherein determining the second metric comprises:

determining a first color difference signal by subtracting a first color component of the image from a second color component of the image;

determining a second color difference signal by subtracting a third color component of the image from the first color component of the image; and

calculating a ratio of a first standard deviation of the first color difference signal to a second standard deviation of the second color difference signal to determine the second metric.

9. The method of claim 1 wherein the plurality of metrics includes a second metric based on an amount of specular reflection in the image.

10. The method of claim 1 wherein the plurality of metrics includes a second metric based on an amount of glare in the image.

11. The method of claim 1 wherein the plurality of metrics includes a second metric based on a number and types of occlusions in the image.

12. The method of claim 1 wherein the plurality of metrics includes a second metric based on a gaze angle of an eye depicted in the image.

13. The method of claim 1 wherein the plurality of metrics includes a second metric based on the segmentation quality of a sclera depicted in the image.

14. A system comprising:

one or more data processing apparatus programmed to perform operations comprising:

obtaining a plurality of images of an eye, wherein the images each include a view of a portion of a vasculature of the eye;

determining a plurality of respective metrics for each of the images based on analyzing image data elements of the image, wherein the respective metrics include a first metric that represents a calculated area of one or more connected structures of programmatically enhanced vasculature in the image and a second metric that represents an extent of detected vasculature in the image based on a plurality of color difference signals determined from color components of the image;

calculating a respective first score for each of the images wherein the first score represents a comparison of the respective metrics of the image with corresponding metrics for a reference image;

calculating a second score based on, at least, a combination of the first scores; and

granting a user access to a device or a service based on the second score.

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

before calculating the respective first score for each of the images, determining a respective quality score for each of the images based on the metrics, wherein the quality score is a prediction of the second score assuming the image and the reference image include a view of a same person's vasculature.

16. The system of claim 15 , wherein the operations further comprise:

accepting or rejecting the image based on the respective quality score.

17. The system of claim 14 wherein calculating the second score based on, at least, the combination of the first scores comprises:

determining a respective weight for each of the first scores; and

combining the first scores weighted by their respective weights to determine the second score.

18. The system of claim 14 , wherein calculating a respective first score for each of the images comprises:

a distance between the metrics for the image and the metrics for the reference image.

19. The system of claim 18 wherein the distance is a Euclidean distance, a correlation coefficient, a modified Hausdorff distance, a Mahalanobis distance, a Bregman divergence, a cosine similarity, a Kullback-Leibler distance, or a Jensen-Shannon divergence.

20. The system of claim 14 wherein determining the first metric comprises:

dilating the vasculature in the image so that one or more disconnected veins in the vasculature become connected;

thinning the dilated vasculature in the image; and

determining a portion of the image that contains the thinned dilated vasculature.

21. The system of claim 14 wherein the second metric is for comparing an extent of eye vasculature detected across different color components in the image, and wherein determining the second metric comprises:

determining a first color difference signal by subtracting a first color component of the image from a second color component of the image;

determining a second color difference signal by subtracting a third color component of the image from the first color component of the image; and

calculating a ratio of a first standard deviation of the first color difference signal to a second standard deviation of the second color difference signal to determine the second metric.

22. The system of claim 14 wherein the plurality of metrics includes a second metric based on an amount of specular reflection in the image.

23. The system of claim 14 wherein the plurality of metrics includes a second metric based on an amount of glare in the image.

24. The system of claim 14 wherein the plurality of metrics includes a second metric based on a number and types of occlusions in the image.

25. The system of claim 14 wherein the plurality of metrics includes a second metric based on a gaze angle of an eye depicted in the image.

26. The system of claim 14 wherein the plurality of metrics includes a second metric based on the segmentation quality of a sclera depicted in the image.

27. A non-transitory computer-readable storage medium having instructions stored thereon that, when executed by data processing apparatus, cause the data processing apparatus to perform operations comprising:

obtaining a plurality of images of an eye, wherein the images each include a view of a portion of a vasculature of the eye;

determining a plurality of respective metrics for each of the images based on analyzing image data elements of the image, wherein the respective metrics include a first metric that represents a calculated area of one or more connected structures of programmatically enhanced vasculature in the image and a second metric that represents an extent of detected vasculature in the image based on a plurality of color difference signals determined from color components of the image;

calculating a respective first score for each of the images wherein the first score represents a comparison of the respective metrics of the image with corresponding metrics for a reference image;

calculating a second score based on, at least, a combination of the first scores; and

granting a user access to a device or a service based on the second score.

28. The computer-readable storage medium of claim 27 , wherein the operations further comprise:

before calculating the respective first score for each of the images, determining a respective quality score for each of the images based on the metrics, wherein the quality score is a prediction of the second score assuming the image and the reference image include a view of a same person's vasculature.

29. The computer-readable storage medium of claim 28 , wherein the operations further comprise:

accepting or rejecting the image based on the respective quality score.

30. The computer-readable storage medium of claim 27 wherein calculating the second score based on, at least, the combination of the first scores comprises:

determining a respective weight for each of the first scores; and

combining the first scores weighted by their respective weights to determine the second score.