Methods, systems, and media for detecting gaze locking

1. A method for detecting gaze locking, the method comprising:

receiving an input image including a face;

locating a pair of eyes in the face of the input image;

generating a first coordinate frame based on one of the pair of eyes and generating a second coordinate frame based on another of the pair of eyes;

concatenating the first coordinate frame and the second coordinate frame to produce a concatenated coordinate frame;

identifying an eye region in the concatenated coordinate frame;

generating, using a hardware processor, a feature vector by concatenating values of pixels in the eye region; and

determining, based on the feature vector, whether the face is gaze locking.

2. The method of claim 1 , further comprising:

identifying a set of fiducial points in the input image; and

locating the pair of eyes in the input image based on the set of fiducial points.

3. The method of claim 2 , further comprising identifying the eye region in the concatenated coordinate frame based on the set of fiducial points.

4. The method of claim 1 , further comprising applying a first affine transformation to a portion of the input image corresponding to the pair of eyes to produce the first coordinate frame.

5. The method of claim 1 , further comprising concatenating intensity values of the pixels in the eye region to produce the feature vector.

6. The method of claim 1 , further comprising applying a mask to the concatenated coordinate frame to produce a masked image.

7. The method of claim 6 , further comprising identifying the eye region based on the masked image.

8. The method of claim 6 , further comprising:

concatenating intensity values of pixels in the masked image to produce the feature vector;

normalizing the feature vector to produce a normalized feature vector; and

determining whether the face is gaze locking based on the normalized feature vector.

9. The method of claim 1 , further comprising:

compressing the feature vector to generate a compressed feature vector; and

determining, based on the compressed feature vector, whether the face is gaze locking using a classifier.

10. The method of claim 9 , wherein the classifier is a binary classifier.

11. A system for detecting gaze locking, the system comprising:

at least one hardware processor that is configured to:

receive an input image including a face;

locate a pair of eyes in the face of the input image;

generate a first coordinate frame based on one of the pair of eyes and generate a second coordinate frame based on another of the pair of eyes;

concatenate the first coordinate frame and the second coordinate frame to produce a concatenated coordinate frame;

identify an eye region in the concatenated coordinate frame;

generate a feature vector by concatenating values of pixels in the eye region; and

determine, based on the feature vector, whether the face is gaze locking.

12. The system of claim 11 , wherein the hardware processor is further configured to:

identify a set of fiducial points in the input image; and

locate the pair of eyes in the input image based on the set of fiducial points.

13. The system of claim 12 , wherein the hardware processor is further configured to identify the eye region in the concatenated coordinate frame based on the set of fiducial points.

14. The system of claim 11 , wherein the hardware processor is further configured to apply a first affine transformation to a portion of the input image corresponding to the pair of eyes to produce the first coordinate frame.

15. The system of claim 11 , wherein the hardware processor is further configured to concatenate intensity values of the pixels in the eye region to produce the feature vector.

16. The system of claim 11 , wherein the hardware processor is further configured to apply a mask to the concatenated coordinate frame to produce a masked image.

17. The system of claim 16 , wherein the hardware processor is further configured to identify the eye region based on the masked image.

18. The system of claim 16 , wherein the hardware processor is further configured to:

concatenate intensity values of pixels in the masked image to produce the feature vector;

normalize the feature vector to produce a normalized feature vector; and

determine whether the face is gaze locking based on the normalized feature vector.

19. The system of claim 11 , wherein the hardware processor is further configured to:

compress the feature vector to generate a compressed feature vector; and

determine, based on the compressed feature vector, whether the face is gaze locking using a classifier.

20. The system of claim 19 , wherein the classifier is a binary classifier.

21. A non-transitory computer-readable medium containing computer-executable instructions that, when executed by a processor, cause the processor to perform a method for detecting gaze locking, the method comprising:

receiving an input image including a face;

locating a pair of eyes in the face of the input image;

generating a first coordinate frame based on one of the pair of eyes and generating a second coordinate frame based on another of the pair of eyes;

concatenating the first coordinate frame and the second coordinate frame to produce a concatenated coordinate frame;

identifying an eye region in the concatenated coordinate frame;

generating a feature vector by concatenating values of pixels in the eye region; and

determining, based on the feature vector, whether the face is gaze locking.

22. The non-transitory computer-readable medium of claim 21 , wherein the method further comprises:

identifying a set of fiducial points in the input image; and

locating the pair of eyes in the input image based on the set of fiducial points.

23. The non-transitory computer-readable medium of claim 22 , wherein the method further comprises identifying the eye region in the concatenated coordinate frame based on the set of fiducial points.

24. The non-transitory computer-readable medium of claim 21 , wherein the method further comprises applying a first affine transformation to a portion of the input image corresponding to the pair of eyes to produce the first coordinate frame.

25. The non-transitory computer-readable medium of claim 21 , wherein the method further comprises concatenating intensity values of the pixels in the eye region to produce the feature vector.

26. The non-transitory computer-readable medium of claim 21 , wherein the method further comprises applying a mask to the concatenated coordinate frame to produce a masked image.

27. The non-transitory computer-readable medium of claim 26 , wherein the method further comprises identifying the eye region based on the masked image.

28. The non-transitory computer-readable medium of claim 26 , wherein the method further comprises:

concatenating intensity values of pixels in the masked image to produce the feature vector;

normalizing the feature vector to produce a normalized feature vector; and

determining whether the face is gaze locking based on the normalized feature vector.

29. The non-transitory computer-readable medium of claim 21 , wherein the method further comprises:

compressing the feature vector to generate a compressed feature vector; and

determining, based on the compressed feature vector, whether the face is gaze locking using a classifier.

30. The non-transitory computer-readable medium of claim 29 , wherein the classifier is a binary classifier.