Iris Recognition Using Local and Global Iris Image Moment Features

Iris recognition in a less constrained environment is challenging. In less constrained environment, iris is subject to rotation, scale, translation, and illumination changes. Therefore it is necessary to extract and represent relevant iris texture features. In this paper, we propose a hybrid method for iris recognition using Zernike moment, Maitra's moment. Iris features are represented using moments. These moments are invariant to scale, rotation, translation and contrast. Global iris texture features are represented using Zernike moments. Normalized iris images is partitioned into non overlapping blocks. For each block, local features are represented using HU and Maitra's moment invariants. Finally global and local features are concatenated to represent iris feature vector. K-Nearest Neighbor (KNN) is used as a classifier. Experiments with publicly available databases: CASIA.v4.interval, PolyU, and IITD iris database shows good recognition accuracy.