Spoof detection via 3D reconstruction

1. A method comprising:

receiving, at one or more processing devices, a first sequence of images of a first subject captured by an image acquisition device over a time period in which relative locations of the image acquisition device with respect to the first subject varies;

selecting, by the one or more processing devices, a first image and a second image from the first sequence of images, wherein the first image represents a first relative location of the image acquisition device with respect to the first subject, and the second image represents a second relative location, different from the first relative location, of the image acquisition device with respect to the first subject;

generating, by the one or more processing devices and using the first image and the second image as a stereo pair, a first three-dimensional representation of the first subject;

determining, based on the first three-dimensional representation, that the first subject is an alternative representation of a live person;

responsive to determining that the first subject is an alternative representation of a live person, preventing access to a secure system;

receiving, at the one or more processing devices, a second sequence of images of a second subject captured by the image acquisition device over a time period in which relative locations of the image acquisition device with respect to the second subject varies;

selecting, by the one or more processing devices, a third image and a fourth image from the second sequence of images, wherein the third image represents a first relative location of the image acquisition device with respect to the second subject, and the fourth image represents a second relative location, different from the first relative location, of the image acquisition device with respect to the second subject;

generating, by the one or more processing devices and using the third image and the fourth image as a stereo pair, a second three-dimensional representation of the second subject;

determining, based on the second three-dimensional representation, that the second subject is a live person; and

responsive to determining that the second subject is a live person, initiating an authentication process for determining if the live person is authorized to access the secure system.

2. The method of claim 1 , wherein the image acquisition device is a camera disposed on a mobile device, and the first sequence of images are captured as the mobile device is moved towards the first subject.

3. The method of claim 1 , wherein the image acquisition device is a camera disposed on a kiosk device, and the first sequence of images are captured as the first subject moves towards the kiosk device.

4. The method of claim 1 , wherein the alternative representation of the live person comprises a photograph of the live person printed on paper, or presented on a display device.

5. The method of claim 1 , wherein the first three-dimensional representation is generated using a photometric stereo process for estimating a three-dimensional shape from two-dimensional images.

6. The method of claim 5 , wherein the photometric stereo process comprises a structure-from-motion process.

7. The method of claim 5 , wherein the photometric stereo process comprises a shape-from-shading process.

8. An system comprising:

an image acquisition device that captures a first sequence of images of a first subject over a time period in which relative locations of the image acquisition device with respect to the first subject varies; and

an image analysis engine comprising one or more processing devices, wherein the image analysis engine:

receives the first sequence of images,

selects a first image and a second image from the first sequence of images, wherein the first image represents a first relative location of the image acquisition device with respect to the first subject, and the second image represents a second relative location, different from the first relative location, of the image acquisition device with respect to the first subject,

generates, using the first image and the second image as a stereo pair, a first three-dimensional representation of the first subject,

determines, based on the first three-dimensional representation, that the first subject is an alternative representation of a live person,

responsive to determining that the first subject is an alternative representation of a live person, prevents access to a secure system;

receives a second sequence of images of a second subject captured by the image acquisition device over a time period in which relative locations of the image acquisition device with respect to the second subject varies;

selects a third image and a fourth image from the second sequence of images, wherein the third image represents a first relative location of the image acquisition device with respect to the second subject, and the fourth image represents a second relative location, different from the first relative location, of the image acquisition device with respect to the second subject;

generates, using the third image and the fourth image as a stereo pair, a second three-dimensional representation of the second subject;

determines, based on the second three-dimensional representation, that the second subject is a live person; and

responsive to determining that the second subject is a live person, initiates an authentication process for determining if the live person is authorized to access the secure system.

9. The system of claim 8 , wherein the image acquisition device comprises a camera disposed on a mobile device, and the first sequence of images are captured as the mobile device is moved towards the first subject.

10. The system of claim 8 , wherein the image acquisition device is a camera disposed on a kiosk device, and the first sequence of images are captured as the first subject moves towards the kiosk device.

11. The system of claim 8 , wherein the alternative representation of the live person comprises a photograph of the live person printed on paper, or presented on a display device.

12. The system of claim 8 , wherein the image analysis engine generates the first three-dimensional representation using a photometric stereo process for estimating a three-dimensional shape from two-dimensional images.

13. The system of claim 12 , wherein the photometric stereo process comprises a structure-from-motion process.

14. The system of claim 12 , wherein the photometric stereo process comprises a shape-from-shading process.

15. One or more non-transitory machine-readable storage devices having encoded thereon computer readable instructions for causing one or more processing devices to perform operations comprising:

obtaining a first sequence of images of a first subject captured by an image acquisition device over a time period in which relative locations of the image acquisition device with respect to the first subject varies;

selecting a first image and a second image from the first sequence of images, wherein the first image represents a first relative location of the image acquisition device with respect to the first subject, and the second image represents a second relative location, different from the first relative location, of the image acquisition device with respect to the first subject;

generating, using the first image and the second image as a stereo pair, a first three-dimensional representation of the first subject;

determining, based on the first three-dimensional representation, that the first subject is an alternative representation of a live person;

responsive to determining that the first subject is an alternative representation of a live person, preventing access to a secure system;

obtaining a second sequence of images of a second subject captured by the image acquisition device over a time period in which relative locations of the image acquisition device with respect to the second subject varies;

selecting a third image and a fourth image from the second sequence of images, wherein the third image represents a first relative location of the image acquisition device with respect to the second subject, and the fourth image represents a second relative location, different from the first relative location, of the image acquisition device with respect to the second subject;

generating using the third image and the fourth image as a stereo pair, a second three-dimensional representation of the second subject;

determining, based on the second three-dimensional representation, that the second subject is a live person; and

responsive to determining that the second subject is a live person, initiating an authentication process for determining if the live person is authorized to access the secure system.

16. The one or more non-transitory machine-readable storage devices of claim 15 , wherein the image acquisition device is a camera disposed on a mobile device, and the first sequence of images are captured as the mobile device is moved towards the first subject.

17. The one or more non-transitory machine-readable storage devices of claim 15 , wherein the image acquisition device is a camera disposed on a kiosk device, and the first sequence of images are captured as the first subject moves towards the kiosk device.