Iris image analysis for biometric system
Uttam M. Chaskar, Mukul Sutaone, S. Panwar · 2009
Iris recognition is accepted as a one of the most efficient biometric methods. Practical implementation of this biometric method requires special image processing, where the iris feature extraction plays a crucial role. In this paper, we analyze the effect of various parameters of algorithms, different image acquisition system specifications, various environment, and different size of images for iris recognition system. Speed, accuracy and acceptability of iris recognition system depend on all the above parameters. Work has been carried out on various eye databases such as UPOL, CASIA, UBIRIS, and some of the images of eyes taken by our image acquisition system. In iris recognition process maximum time is required for localization. It has been found that good contrast image has more accuracy of localization and small size image results in fast localization. Speed of recognition system also depends upon the specifications of data acquisition system. FAR, FRR, speed and performance of overall recognition system also depends upon iris localization, size of normalized iris and encoding filter parameters. All these factors have been found variable, depending upon data acquisition used. (7 pages)