Polar binary encoding of iris texture for real-time personal identification

Over the last two decades human iris has proven to be the most reliable and accurate biometric. However, iris recognition faces various challenges because of large template size and computationally expensive encoding schemes. In this paper, a novel polar domain iris normalization model is proposed by encoding and matching iris images using simple binarization technique for real time personal identification. To determine the performance of the proposed normalization and encoding scheme, a comparative analysis is presented with other encoding techniques on the basis of equal error rates and template size. The comparison is performed on standard iris databases. It has been evaluated that the proposed technique ideally suits real time identification, as it reduces both error rates and template size in comparison to other encoding schemes presented in literature.