1. Segmentation of iris using daugman’s algorithm
Authors : Patil Ashwini B, Dr.mrs.v.jayashree
Pages : 243-251
DOI : http://dx.doi.org/10.21172/1.102.39
Keywords : Canny Edge Detector, Hough Transform, Daugman’s rubber sheet model, Image Normalization. Abstract :Security is the main concern in majority of the recognition techniques that are used in various Biometric systems. The focus here is to subject the eye images to segmentation process to reduce the redundant data and noise from the image of eye. So this paper represents the segmentation of the eye images to extract the actual iris region from entire eye image. The eye images are taken from the CASIA database (Chinese academy of Sciences). All The iris images were subjected to segmentation process using canny edge detector for edge detection from which outer and inner boundaries of iris eye images were obtained by Hough transform. This was followed by eyelashes occlusion and normalization by subjecting these images to Daugman’s rubber sheet model which then generated the iris templates. From the normalized 280x320 pixel image, redundant data was removed converting the iris image to a size of 50x250 pixels. This procedure was applied on all 756 eye images of 108 persons each with 7 different orientations. The experimentation has shown that, there is a large variation in average value of iris/pupil radius (from 100 to 128.53 pixels)/ (38.71 to 45.72) respectively whereas variation in standard deviation value for iris/pupil (from 1.13 to 5.77) / (from1.25 to 3.4) respectively for the different iris images of 10 people. This is a desired feature of iris segmentation for iris recognition. Thus this segmentation approach has resulted into reduced useful iris information for further classification and iris identification.
Citing this Journal Article :Patil Ashwini B, Dr.mrs.v.jayashree, "Segmentation of iris using daugman’s algorithm", Volume 10 Issue 2 - April 2018, 243-251
Click here to Submit Copyright Takedown Notice for this article.