An efficient iris code storing and searching technique for Iris Recognition using non-homogeneous K-d tree

Iris Recognition is increasingly being used as the main method for biometric authentication since it is highly reliable and accurate. It considers the unique patterns of the iris to identify personnel by applying pattern recognition techniques. As the usage of iris recognition system increases, the number of iris code to be stored and retrieved for matching in large database of irises increases proportionally. Searching for the iris match in a huge database of iris templates poses challenges to research community in terms of retrieval accuracy and efficiency. The non-homogenous K-d tree structure used in the proposed model stores and matches iris code, which improves the search accuracy of the iris recognition system. The proposed model is tested on IITD and CASIA datasets.