Detecting irises and pupils in images of humans
First Claim
1. An eye detection method for use with a digital image having a plurality of pixels, said method comprising the steps of:
- locating a plurality of candidate eyes within the image;
finding pixels of said digital image having a predetermined characteristic and a predetermined association to respective said eyes, to provide pixels of interest;
grouping said pixels of interest associated with each of said eyes into a plurality of different groups;
fitting parametric boundaries on each of said groups to define a plurality of detected eye features, said boundaries having a predefined uniform shape and a size matched to a respective said group;
scoring each of said detected eye features as to a geometric relationship between the respective one of said boundaries and said pixels of interest associated with the respective said eye to provide respective eye feature scores.
5 Assignments
0 Petitions
Accused Products
Abstract
In an eye detection method, a plurality of candidate eyes are located within a digital image. Pixels of interest having a predetermined characteristic and a predetermined association to respective eyes are found. Pixels of interest associated with each eye are grouped. Parametric boundaries are fit on each of the groups to define a plurality of detected eye features. The boundaries have a predefined uniform shape and a size matched to a respective group. Each of the detected eye features is scored as to a geometric relationship between the respective boundary and pixels of interest associated with the respective eye to provide eye feature scores.
136 Citations
64 Claims
-
1. An eye detection method for use with a digital image having a plurality of pixels, said method comprising the steps of:
-
locating a plurality of candidate eyes within the image;
finding pixels of said digital image having a predetermined characteristic and a predetermined association to respective said eyes, to provide pixels of interest;
grouping said pixels of interest associated with each of said eyes into a plurality of different groups;
fitting parametric boundaries on each of said groups to define a plurality of detected eye features, said boundaries having a predefined uniform shape and a size matched to a respective said group;
scoring each of said detected eye features as to a geometric relationship between the respective one of said boundaries and said pixels of interest associated with the respective said eye to provide respective eye feature scores. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. An eye detection method for use with a digital image having a plurality of pixels, said method comprising the steps of:
-
locating one or more candidate eyes within the image;
identifying one or more local neighborhoods, each said local neighborhood being inclusive of and larger than a respective one of said candidate eyes;
determining a plurality of different groups of pixels of interest within each of said local neighborhoods;
fitting predefined parametric boundaries on each of said groups to define a plurality of detected eye features;
scoring each of said detected eye features as to a geometric relationship between pixels of interest in the local neighborhood and the respective said boundary to provide respective eye feature scores. - View Dependent Claims (19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50)
-
-
51. An eye detection method for use with a digital image having a plurality of pixels, said method comprising the steps of:
-
locating one or more redeye defect pairs in the digital image;
identifying a local neighborhood relative to each member of each of said redeye defect pairs;
ascertaining pixels of interest within each said local neighborhood, said pixels of interest each having a predetermined characteristic;
determining a circle best fitting said pixels of interest in each said local neighborhood. - View Dependent Claims (52, 53, 54, 55, 56, 57, 58)
-
-
59. A method of detecting a human iris in a digital image having a plurality of pixels and one or more redeye defects, said method comprising:
-
locating one or more redeye defect pairs in the digital image;
identifying a local neighborhood relative to each member of each of said redeye defect pairs, said local neighborhoods each having a size that is a function of the separation of members of the respective said redeye defect pair;
producing an edge map of each of said local neighborhoods;
identifying positions on each said edge map having an edge magnitude greater than a predefined threshold to define pixels of interest;
grouping said pixels of interest associated with each of said redeye defects into a plurality of different groups, each group having two pixels;
fitting circles on each of said groups, wherein said edge directions are tangent to the respective said circle at each of said pixels of each said group, to define a plurality of detected eye features;
counting a number of said pixels of interest within a predetermined distance of each of said circles; and
assigning a weight to each of said pixels of interest within said predetermined distance of respective said circles.
-
-
60. A method of correcting redeye defects in an image having pixels of a human subject, said method comprising:
-
locating a pair of redeye defects representing a right eye and a left eye, said defects each having one or more pixels;
identifying a local neighborhood associated with each of said eyes;
searching in each said local neighborhood to detect a candidate eye feature having a circular shape and provide a pair of detected eye features;
using said detected eye features and associated said redeye defects to correct the image. - View Dependent Claims (61, 62)
-
-
63. A computer program product for use with a digital image having a plurality of pixels, the computer program product comprising computer readable storage medium having a computer program stored thereon for performing the steps of:
-
locating a plurality of candidate eyes within the image;
finding pixels of said digital image having a predetermined characteristic and a predetermined association to respective said eyes, to provide pixels of interest;
grouping said pixels of interest associated with each of said eyes into a plurality of different groups;
fitting parametric boundaries on each of said groups to define a plurality of detected eye features, said boundaries having a predefined uniform shape and a size matched to a respective said group;
scoring each of said detected eye features as to a geometric relationship between the respective one of said boundaries and said pixels of interest associated with the respective said eye to provide respective eye feature scores.
-
-
64. An eye detection apparatus for use with a digital image having a plurality of pixels, comprising:
-
means for locating a plurality of candidate eyes within the image;
means for finding pixels of said digital image having a predetermined characteristic and a predetermined association to respective said eyes, to provide pixels of interest;
means for grouping said pixels of interest associated with each of said eyes into a plurality of different groups;
means for fitting parametric boundaries on each of said groups to define a plurality of detected eye features, said boundaries having a predefined uniform shape and a size matched to a respective said group;
means for scoring each of said detected eye features as to a geometric relationship between the respective one of said boundaries and said pixels of interest associated with the respective said eye to provide respective eye feature scores.
-
Specification