Spoof detection using intraocular reflection correspondences
Methods, systems, and computer-readable storage media for determining that a subject is a live person include capturing one or more images of two eyes of a subject. The one or more images, from each of the two eyes are used to obtain respective corneal reflections. Depth information associated with a scene in front of the subject is determined, based on an offset between the respective corneal reflections. A determination is made, based at least on the depth information, that the subject is a live person. Responsive to determining that the subject is a live person, an authentication process is initiated to authenticate the subject.
1. A computer-implemented method comprising:
capturing one or more images of two eyes of a subject;
obtaining, from the one or more images, respective corneal reflections from each of the two eyes;
determining, based on an offset between the respective corneal reflections, depth information associated with a scene in front of the subject;
determining a scene indicator representative of at least one characteristic of the scene in front of the subject, wherein the scene indicator comprises a pose of the subject and wherein the pose of the subject comprises an elbow joint angle, a wrist joint angle, and a shoulder joint angle;
determining, based at least on the depth information, that the subject is a live person; and
in response to determining that the subject is a live person, initiating an authentication process to authenticate the subject.
2. The computer-implemented method of claim 1 , wherein the scene indicator further comprises a content of a display panel of the user device, or an orientation of the user device.
3. The computer-implemented method of claim 1 , further comprising:
comparing the scene indicator to a reference scene indicator.
4. The computer-implemented method of claim 3 , further comprising:
determining a difference between the scene indicator and the reference scene indicator; and
comparing the difference to a threshold.
5. The computer-implemented method of claim 1 , wherein determining that the subject is the live person comprises processing the scene indicator using a machine learning process trained to discriminate between scene indicators of live persons and artificial scene indicators of alternative representations of the live persons.
6. The computer-implemented method of claim 1 , wherein generating the depth information comprises:
generating, based on the offset, a stereo correspondence between the corneal reflections;
generating rectified images by rectifying the one or more images based on information on corneal curvature; and
generating a depth map from the rectified images based on the stereo correspondence.
7. A non-transitory, computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
capturing one or more images of two eyes of a subject;
obtaining, from the one or more images, respective corneal reflections from each of the two eyes;
determining, based on an offset between the respective corneal reflections, depth information associated with a scene in front of the subject;
determining a scene indicator representative of at least one characteristic of the scene in front of the subject, wherein the scene indicator comprises a pose of the subject and wherein the pose of the subject comprises an elbow joint angle, a wrist joint angle, and a shoulder joint angle;
determining, based at least on the depth information, that the subject is a live person; and
in response to determining that the subject is a live person, initiating an authentication process to authenticate the subject.
8. The non-transitory, computer-readable medium of claim 7 , wherein the scene indicator further comprises a content of a display panel of the user device, or an orientation of the user device.
9. The non-transitory, computer-readable medium of claim 7 , wherein the operations further comprise:
comparing the scene indicator to a reference scene indicator.
10. The non-transitory, computer-readable medium of claim 9 , wherein the operations further comprise:
determining a difference between the scene indicator and the reference scene indicator; and
comparing the difference to a threshold.
11. The non-transitory, computer-readable medium of claim 7 , wherein determining that the subject is the live person comprises processing the scene indicator using a machine learning process trained to discriminate between scene indicators of live persons and artificial scene indicators of alternative representations of the live persons.
12. The non-transitory, computer-readable medium of claim 7 , wherein generating the depth information comprises:
generating, based on the offset, a stereo correspondence between the corneal reflections;
generating rectified images by rectifying the one or more images based on information on corneal curvature; and
generating a depth map from the rectified images based on the stereo correspondence.
13. A computer-implemented system, comprising:
one or more computers; and
one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing one or more instructions that, when executed by the one or more computers, perform operations comprising:
capturing one or more images of two eyes of a subject;
obtaining, from the one or more images, respective corneal reflections from each of the two eyes;
determining, based on an offset between the respective corneal reflections, depth information associated with a scene in front of the subject;
determining a scene indicator representative of at least one characteristic of the scene in front of the subject, wherein the scene indicator comprises a pose of the subject and wherein the pose of the subject comprises an elbow joint angle, a wrist joint angle, and a shoulder joint angle;
determining, based at least on the depth information, that the subject is a live person; and
in response to determining that the subject is a live person, initiating an authentication process to authenticate the subject.
14. The computer-implemented system of claim 13 , wherein the scene indicator further comprises a content of a display panel of the user device, or an orientation of the user device.
15. The computer-implemented system of claim 13 , wherein the operations further comprise:
comparing the scene indicator to a reference scene indicator;
determining a difference between the scene indicator and the reference scene indicator; and
comparing the difference to a threshold.
16. The computer-implemented system of claim 13 , wherein determining that the subject is the live person comprises processing the scene indicator using a machine learning process trained to discriminate between scene indicators of live persons and artificial scene indicators of alternative representations of the live persons.
17. The computer-implemented system of claim 13 , wherein generating the depth information comprises:
generating, based on the offset, a stereo correspondence between the corneal reflections;
generating rectified images by rectifying the one or more images based on information on corneal curvature; and
generating a depth map from the rectified images based on the stereo correspondence.