DocumentCode :
2655537
Title :
Non-circular pupil localization in iris images
Author :
Basit, A. ; Javed, M.Y. ; Masood, S.
Author_Institution :
Dept. of Comput. Eng., Nat. Univ. of Sci. & Technol. (NUST), Rawalpindi
fYear :
2008
fDate :
18-19 Oct. 2008
Firstpage :
228
Lastpage :
231
Abstract :
Iris localization is the most important step in iris recognition systems, due to the fact that all the subsequent steps, iris normalization, feature extraction and matching, depends on its accuracy. Iris localization means segmentation of iris from the other parts of eye like pupil, sclera, eyelids and eyelashes in the image of an eye. A new method of exact pupil localization is proposed in this paper. Pupil is localized by finding a point in the pupil and then center is obtained using centroid of the pupil region whereas radius for further processing is calculated from the binary image of the region. Non-circular boundary of the pupil is ciphered by divide and conquer rule. Circular boundary of pupil is divided into a specified number of points. These points are repositioned with respect to the maximum gradient and then joined together to obtained exact boundary of the pupil. Iris outer boundary and eyelids are determined using intensity gradient. Experimental results show that the proposed method of iris localization is quite effective.
Keywords :
biometrics (access control); edge detection; feature extraction; image matching; image recognition; image segmentation; binary image; divide-and-conquer rule; feature extraction; image matching; image segmentation; intensity gradient determination; iris images; iris recognition system; noncircular pupil localization; pupil circular boundary extraction; pupil region centroid; Biometrics; Detectors; Educational institutions; Eyelashes; Eyelids; Humans; Image edge detection; Image segmentation; Iris recognition; Mechanical engineering; Biometrics; Image Segmentation; Iris localization; Pupil localization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Emerging Technologies, 2008. ICET 2008. 4th International Conference on
Conference_Location :
Rawalpindi
Print_ISBN :
978-1-4244-2210-4
Electronic_ISBN :
978-1-4244-2211-1
Type :
conf
DOI :
10.1109/ICET.2008.4777505
Filename :
4777505
Link To Document :
بازگشت