Implementation of Iris Recognition on FPGA

Iris recognition is accepted as one of the most efficient biometric method. Implementing this method to the practical system requires specific image processing where the iris feature extraction plays a crucial role. The first step in recognition is iris localization which consists in finding the iris boundaries as well as eyelids. In iris recognition system the maximum time is required for localization. In the recognition process the image acquisition system and size of theimage is an important for accuracy and overall system performance. Most of image processing algorithms are computationally intensive, so it is desirable to implement them in high performance reconfigurable systems. Recently, Field Programmable Gate Array (FPGA) technology is appropriate target for the implementation of algorithms for image processing. In this paper iris localization and iris recognition algorithms are implemented in MATLAB first, then the attempt has been made to design and develop iris recognition algorithm on FPGA. In addition, some of the optimizations are proposed and analyzed. Proposed solution is applied to 150×200 size colored images acquired under realistic conditions (UBIRIS Database).