Preprocessing Algorithm For Iris Recognition

In recent years, iris recognition has aroused much concern in academic and industrial fields, due to the unique physiological characteristics of iris in biometric feature recognition. Nowadays, how to quickly and precisely make a real-time and automatic iris identification or verification is still a problem. Nevertheless, to develop an iris-image-preprocessing model may provide a basis for coping with this problem. With respect to the existing algorithms of preprocessing assessment on iris images, most of them based on the sequence of images, prolonging additional time of iris image processing.Through analyzing existing algorithms, An new preprocessing assessment algorithm is presented in the thesis on iris image in real-time iris recognition system, adopting a step-by-step and real-time procedure to process iris image. The main tasks of this thesis are as follows:1. Assessing the quality of iris imageThe first is to analyze all kinds of factors which affect the quality of iris image, The second is to set corresponding rules and criteria referring to different factors. The new quality algorithm solve image definition and eyelash occlusion problem.2. Locating irisIris localization algorithm integrate with the result of quality assessment, The first is to select interest regions in iris image, the second is to set a binary threshold of pupil by adopting the method of Maximum Variance between Clusters; the third is to seek the outer edge points on the basis of gradient value of row-pixels in the selected regions, and lastly to locate the inner and outer edges of iris by means of curve fitting. Therefore, this algorithm avoids iterative calculation and shortens the time of preprocessing iris images.3. Establishing the assessment model of real-time preprocessingThe first is to making correlation analysis of all the kinds of factors which affect the quality of iris images step by step. The second is to make real-time judgment and assessment on the image quality. The last is to establish the preprocessing assessment model of real-time iris recognition system based on PC platforms.The algorithms put forward in this thesis have been validated by a plenty of simulating tests and experiments, and the experimental results show that these algorithms are feasible to make rapid and precise preprocessing iris images.