Image acquisition system for off-axis eye images
An image acquisition system determines first and second sets of points defining an iris-pupil boundary and an iris-sclera boundary in an acquired image; determines respective ellipses fitting the first and second sets of points; determines a transformation to transform one of the ellipses into a circle on a corresponding plane; using the determined transformation, transforms the selected ellipse into a circle on the plane; using the determined transformation, transforms the other ellipse into a transformed ellipse on the plane; determines a plurality of ellipses on the plane for defining an iris grid, by interpolating a plurality of ellipses between the circle and the transformed ellipse; moves the determined grid ellipses onto the iris in the image using an inverse transformation of the determined transformation; and extracts an iris texture by unwrapping the iris and interpolating image pixel values at each grid point defined along each of the grid ellipses.
1. A method comprising:
acquiring, through an image sensor, an image including an iris of an eye of a subject;
determining a set of first points defining an iris-pupil boundary in the image;
determining a set of second points defining an iris-sclera boundary in the image;
determining a first ellipse fitting the set of first points;
determining a second ellipse fitting the set of second points;
transforming a selected one of the first ellipse or the second ellipse into a circle on a plane;
transforming a non-selected one of the first ellipse or the second ellipse into a transformed ellipse on the plane;
interpolating a plurality of ellipses between (i) the circle and (ii) the transformed ellipse to provide grid ellipses;
moving the grid ellipses onto the iris in the image using an inverse transformation of a transformation related to transforming the first ellipse and the second ellipse into the circle and the transformed ellipse; and
unwrapping the iris and interpolating image pixel values at individual grid points defined along the grid ellipses to extract an iris texture.
2. The method of claim 1 , further comprising:
transforming spatial image coordinates of the first points and the second points into camera coordinates using a matrix of camera intrinsic parameters,
wherein moving the grid ellipses onto the iris within the image comprises moving the grid ellipses onto the iris within the image using (i) the inverse transformation and (ii) the matrix of camera intrinsic parameters.
3. The method of claim 2 , wherein the matrix of camera intrinsic parameters is obtained by a calibration process to determine correction parameters for compensating image distortions introduced by a camera module.
4. The method of claim 1 , wherein after transforming into the circle and the transformed ellipse, the first points and second points are expressed in homogeneous camera coordinates.
5. The method of claim 1 , further comprising:
determining a transformation related to transforming the first ellipse and the second ellipse into the circle and the transformed ellipse by determining a plurality of candidate transformations for transforming the selected one of the first ellipse or the second ellipse into a circle; and
selecting the candidate transformation which is consistent with a gaze vector of the eye by comparing gaze vectors estimated using the candidate transformations with an estimated center of the eye in the image.
6. The method of claim 1 , further comprising:
determining a transformation related to transforming the first ellipse and the second ellipse into the circle and the transformed ellipse by determining a rotation for transforming an oblique elliptical cone having the selected one of the first ellipse or the second ellipse as a base and a vertex at a focal point of a camera into an oblique circular cone having a circular base and a vertex at the focal point of the camera, wherein an inverse rotation of a normal vector to the circular base of the oblique circular cone corresponds to a gaze vector of the eye.
7. The method of claim 6 , wherein determining the rotation comprises:
applying a first rotation to transform the oblique elliptical cone to a non-oblique elliptical cone; and
applying a second rotation to transform the non-oblique elliptical cone to the oblique circular cone.
8. The method of claim 6 , further comprising:
transforming spatial image coordinates of the first points and the second points into camera coordinates using a matrix of camera intrinsic parameters,
wherein moving the grid ellipses onto the iris in the image comprises moving the grid ellipses onto the iris in the image using an inverse rotation of (i) the rotation and (ii) the matrix of camera intrinsic parameters.
9. The method of claim 1 , wherein:
centers of the grid ellipses change linearly between the centers of (i) the circle and (ii) the transformed ellipse;
minor axes and eccentricity of the grid ellipses change linearly between (i) the circle and (ii) the transformed ellipse; and
inclination angles of the grid ellipses correspond to an inclination angle of the transformed ellipse.
10. The method of claim 1 , further comprising:
selecting an image axis for unwrapping the iris; and
generating a plurality of grid points along each grid ellipse moved onto the iris in the image such that corresponding starting grid points along the grid ellipses lay on the image axis.
11. An image acquisition system configured to:
acquire, through an image sensor, an image including an iris of an eye of a subject;
determine a set of first points defining an iris-pupil boundary in the image;
determine a set of second points defining an iris-sclera boundary in the image;
determine a first ellipse fitting the set of first points;
determine a second ellipse fitting the set of second points;
transform a selected one of the first ellipse or the second ellipse into a circle on a plane;
transform a non-selected one of the first ellipse or the second ellipse into a transformed ellipse on the plane;
interpolating a plurality of ellipses between (i) the circle and (ii) the transformed ellipse to provide grid ellipses;
move the grid ellipses onto the iris in the image using an inverse transformation of a transformation related to transforming the first ellipse and the second ellipse into the circle and the transformed ellipse; and
unwrapping the iris and interpolating image pixel values at individual grid points defined along the grid ellipses to extract an iris texture.
12. The image acquisition system of claim 11 , wherein the image acquisition system is further configured to:
transform spatial image coordinates of the first points and the second points into camera coordinates using a matrix of camera intrinsic parameters; and
move the grid ellipses onto the iris within the image using (i) the inverse transformation and (ii) the matrix of camera intrinsic parameters.
13. The image acquisition system of claim 12 , wherein the matrix of camera intrinsic parameters is obtained by a calibration process to determine correction parameters for compensating image distortions introduced by a camera module.
14. The image acquisition system of claim 11 , wherein after transforming into the circle and the transformed ellipse, the first points and second points are expressed in homogeneous camera coordinates.
15. The image acquisition system of claim 11 , wherein the image acquisition system is further configured to:
determine a transformation related to transforming the first ellipse and the second ellipse into the circle and the transformed ellipse by determining a plurality of candidate transformations for transforming the selected one of the first ellipse or the second ellipse into a circle; and
select the candidate transformation which is consistent with a gaze vector of the eye by comparing gaze vectors estimated using the candidate transformations with an estimated center of the eye in the image.
16. The image acquisition system of claim 11 , wherein the image acquisition system is further configured to determine a transformation related to transforming the first ellipse and the second ellipse into the circle and the transformed ellipse by:
determining a rotation for transforming an oblique elliptical cone having the selected one of the first ellipse or the second ellipse as a base and a vertex at a focal point of a camera into an oblique circular cone having a circular base and a vertex at the focal point of the camera, wherein an inverse rotation of a normal vector to the circular base of the oblique circular cone corresponds to a gaze vector of the eye.
17. The image acquisition system of claim 16 wherein the image acquisition system is further configured to determine the rotation by:
applying a first rotation to transform the oblique elliptical cone to a non-oblique elliptical cone; and
applying a second rotation to transform the non-oblique elliptical cone to the oblique circular cone.
18. The image acquisition system of claim 16 , wherein the image acquisition system is further configured to:
transform spatial image coordinates of the first points and the second points into camera coordinates using a matrix of camera intrinsic parameters; and
move the grid ellipses onto the iris in the image using an inverse rotation of (i) the rotation and (ii) the matrix of camera intrinsic parameters.
19. The image acquisition system of claim 11 , wherein:
centers of the grid ellipses change linearly between the centers of (i) the circle and (ii) the transformed ellipse;
minor axes and eccentricity of the grid ellipses change linearly between (i) the circle and (ii) the transformed ellipse; and
inclination angles of the grid ellipses correspond to an inclination angle of the transformed ellipse.
20. The image acquisition system of claim 11 , wherein the image acquisition system is further configured to:
select an image axis for unwrapping the iris; and
generate a plurality of grid points along each grid ellipse moved onto the iris in the image such that corresponding starting grid points along the grid ellipses lay on the image axis.