Eye detection for eyeglass wearers in iris recognition

An eye detection method for eyeglass wearers in iris recognition is proposed. The eye detection method consists of two steps; eye candidate detection and eye validation. In eye candidate detection, eye candidates are searched by employing a mask with 2×3 sub-blocks on a near infrared eye image to compute the intensity difference between the iris region and the surrounding regions. In eye validation, an eye is verified by searching a pupil region in the eye candidates by employing a mask with 3×3 sub-blocks to compute the intensity difference between the pupil region and its surrounding regions. The experiment result shows that the eye detection rate is higher when a 2×3 mask is employed than when a 3×3 mask is employed in the eye candidate detection for eyeglass wearers.