Robust iris identification system using local descriptors

In all biometrics, iris contains rich texture for person identification and has its distinctive advantages because of its large inter-class variability and low intra-class variability. This research provides a fast and robust iris identification system. In this proposed system for iris recognition Daisy descriptors have been used as features which has been modified considering the context of iris recognition. After feature extraction, a Feature matching algorithm has also been proposed which provides much better results. The performance of proposed iris recognition system has been evaluated through exhaustive experimentation on standard CASIA Iris Lamp Database and the results exhibit the improvement in recognition accuracy.