| With the advent of globalization, the world seemingly has become a smaller place and has interconnected people, thus resulting in a very fast-paced life. In the last ten years, the demand for more faster, efficient, and robust technologies for security and surveillance has been phenomenal. The search for a definitive answer to address the above need leads us to biometrics, which has attained immense popularity in the industry.; Biometrics or biometry is defined as a statistical study of some measurable biological phenomena which is unique to every living entity (humans). Taking a cue from these unique characteristics, biometric technologies can be developed. Of the approximately sixteen prominent biometric technologies commercially deployed, iris recognition is regarded as the most reliable and accurate biometric identification system. Current iris recognition systems use the patented algorithms of Dr. John Daugman, and have been successfully deployed at airports and other commercial establishments.; The work presented in this thesis consists of an alternate approach taken in developing algorithms for automatic segmentation of the pupil and iris, and for analyzing the normalized image pattern with log-Gabor filters as opposed to using the 2D Gabor filters used by Daugman. The performance and accuracy of the recognition system is studied and tested. |
