Method and apparatus for predicting eye position
Provided is an apparatus and method for predicting an eye position of user that may detect eye position coordinates in an image which is recently generated, determine a current prediction velocity of an eye movement based on a velocity of a previous eye movement, and predict the eye position coordinates of the user based on the detected eye position coordinates and the current prediction velocity.
1. A method of predicting an eye position, the method comprising:
receiving a first image photographed by a camera;
obtaining current eye position coordinates x n in the first image by detecting a face of a user in the first image;
calculating a current velocity v n based on the current eye position coordinates x n , a previous eye position coordinates x n−1 in a second image generated before the first image, and a difference between points in time at which the first image and the second image are photographed; and
updating a first history of a velocity of an eye movement by adding the current velocity v n to the first history;
determining a current prediction velocity v p n of the eye movement based on the first history and a previous prediction velocity v p n−1 of the eye movement with respect to the second image; and
predicting current prediction eye position coordinates x p n using the current prediction velocity v p n .
2. The method of claim 1 , wherein the updating comprises updating the first history using a first in first out (FIFO) method.
3. The method of claim 1 , wherein the determining comprises:
determining a prediction mode based on the previous prediction
velocity v p n−1 of the eye movement; and
determining the current prediction velocity v p n by applying a filter corresponding to the prediction mode to the first history.
4. The method of claim 3 , wherein the determining of the prediction mode comprises:
comparing the previous prediction velocity v p n−1 to a threshold value; and
determining a mode to be any one of a fast mode and a slow mode based on a result of the comparing.
5. The method of claim 3 , wherein the determining of the prediction mode comprises:
comparing a difference between the first history and a second history comprising prediction velocities determined with respect to a plurality of images, the second history including the previous prediction velocity v p n−1 ; and
determining a mode to be any one of a normal mode and an eccentric mode based on a result of the comparing.
6. The method of claim 3 , wherein the filter is a bilateral filter, and at least one parameter of the bilateral filter is adaptively determined based on the prediction mode.
7. The method of claim 1 , wherein the first history includes a preset number of pieces of velocity information.
8. The method of claim 1 , wherein the predicting of the current prediction eye position coordinates x p n comprises predicting the current prediction eye position coordinates x p n based on the current eye position coordinates x n , the current prediction velocity v p n , and a system delay value.
9. The method of claim 8 , wherein the system delay value is a difference between a point in time at which the first image is photographed and a point in time at which the current prediction eye position coordinates x p n are determined.
10. The method of claim 1 , further comprising:
photographing the first image.
11. A non-transitory computer-readable medium storing program code that, when executed by a processor, performs the method of claim 1 .
12. An apparatus for predicting an eye position, the apparatus comprising:
a memory configured to store a first history of a velocity of an eye movement; and
a processor configured to:
receive a first image photographed by a camera,
obtain current eye position coordinates x n in the first image by detecting a face of a user in the first image,
calculate a current velocity v n based on the current eye position coordinates x n , a previous eye position coordinates x n−1 in a second image generated before the first image, and a difference between points in time at which the first image and the second image are photographed,
update the first history of a velocity of an eye movement by adding the current velocity v n to the first history,
determine a current prediction velocity v p n of the eye movement based on the first history and a previous prediction velocity v p n−1 of the eye movement with respect to the second image, and
predict current prediction eye position coordinates x p n using the current prediction velocity v p n .
13. The apparatus of claim 12 , wherein the processor is configured to update the first history using a first in first out (FIFO) method.
14. The apparatus of claim 12 , wherein the processor is configured to determine a prediction mode based on the previous prediction velocity v p n−1 of the eye movement and determine the current prediction velocity v p n by applying a filter corresponding to the prediction mode to the first history.
15. The apparatus of claim 14 , wherein the filter is a bilateral filter and a parameter of the bilateral filter is adaptively determined based on the prediction mode.
16. The apparatus of claim 12 , wherein the processor is configured to predict the current prediction eye position coordinates x p n based on the current eye position coordinates x n , the current prediction velocity v p n , and a system delay value.
17. The apparatus of claim 16 , wherein the system delay value is a difference between a point in time at which the first image is photographed and a point in time at which the current prediction eye position coordinates x p n is determined.
18. The apparatus of claim 12 , further comprising:
a camera configured to photograph the first image.