Effectual Human Authentication for Critical Security Applications Using Retinal Images

A robust method of human authentication based on the retinal blood vessel pattern is presented in this paper. This method entails a segmentation process to identify retinal blood vessel pattern, template generation consisting of the bifurcation points in the retina and matching of the intersection points in the template patterns. The number of matched blood vessel intersection points between the two patterns compared is used as a measure of similarity. As Liveness detection is a highly desirable anti-spoofing measure in biometric authentication, it is ensured while acquiring retinal images in realtime. The validity of our approach is verified with experimental results obtained from 603 comparisons made using 303 retinal images from three different publicly available databases, namely DRIVE, VARIA and STARE. We found that the proposed retinal recognition method gives 100%, 96.3% and 91.1% recognition rates respectively for the above databases. To the best of our knowledge, this is the first work that uses a large number of retinal images from different retinal databases for the authentication purpose.