IRIS Recognition Using High Level Features

This paper demonstrates that personal identity authentication through comparison of high level features of iris is very effective. The success of a biometric recognition system depends heavily on its feature representation model for biometric patterns. Accomplishing sensitivity to inter-class differences and at the same time robustness against intra-class variations is very difficult. Many biometric representation schemes have been reported but the above issue remains to be resolved. This paper introduces iris recognition using high level features in an attempt to resolve this issue. Huge feature space can be derived with different parameter settings such as distance, location, scale, orientation and number. Feature selection aimed at accurate and sparse representation of ordinal measures. This paper provides separation between inter classes and intra class robustness. High level features of iris provide simple and fast recognition through small feature set using ordinal feature representation.