Retina based biometric authentication using phase congruency
- PDF / 963,303 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 28 Downloads / 203 Views
ORIGINAL ARTICLE
Retina based biometric authentication using phase congruency M. Islamuddin Ahmed • M. Ashraful Amin Bruce Poon • Hong Yan
•
Received: 31 March 2012 / Accepted: 1 June 2013 Ó Springer-Verlag Berlin Heidelberg 2013
Abstract In the evolving paradigm of biometric security concepts, retinal fundus image based authentication systems is regarded as one of the most precise and secure procedures. Although previously only high security buildings could afford to utilize this technology to control access, presently due to easier availability and affordability of high-tech equipment, building and establishments with lesser security significance are also employing and relishing the benefits of this technology. This paper presents a novel biometric authentication method using retinal fundus image which uses an extremely fast optical disc detection procedure, a new and fast vessel segmentation method based on Phase congruency and some proposed new features. Most relevant literatures reveal sole usage of green channel from RGB image. To investigate suitability of other color spaces, images of luminance (Y) and green (G) channel of RGB and YCbCr color space respectively are experimented with. Phase congruency is used for vessel segmentation which uses Fourier components to detect M. I. Ahmed M. A. Amin B. Poon Computer Vision and Cybernetics Group, CSE, Independent University, Bangladesh (IUB), Dhaka, Bangladesh e-mail: [email protected] B. Poon e-mail: [email protected] M. A. Amin (&) H. Yan Depertment of Electronic Engineering, City University of Hong Kong, Hong Kong, Hong Kong e-mail: [email protected] H. Yan e-mail: [email protected] B. Poon H. Yan School of Electrical and Information Engineering, University of Sydney, Sydney, NSW 2006, Australia
edges and by applying pair threshold values binary image of retinal blood vessel tree is acquired. Three different features are extracted for classification purpose. Two separate experiments are performed, EXP-1 using 18 images from 6 individuals and EXP-2 using 18 (authorized) plus 547 (intruder) images, each from a separate individual. For similarity matching, 2-D Correlation Coefficient measure is used. In EXP-1 and EXP-2, maximum accuracy of 94.44 and 93.4 % were achieved in 12.81 and 12.92 s respectively with YCbCr images. Keywords Biometric authentication Bifurcation point Feature point matching Retina image recognition Optical disc
1 Introduction Regulating and restricting access to crucial locations and sensitive information is a very vital and active research topic in today’s world. Previously regulations to access were and still are achieved by means of possession of items (e.g. identity cards) or possession of knowledge (e.g. passwords). However, effectiveness of these security measures seemed to become inadequate as cases of forgeries became alarmingly easy and recurrent. In current crisis situation, biometric security concept was proposed which proved to be timely, appropriate and most effective. Although retinal blood
Data Loading...
