Imperfect Iris Information for Identity Recognition

Iris recognition need user cooperate, but some factors is uncertainty and unstable between image-capture set and user little, the thing that iris information is imperfect in image captured is ubiquitous. According to the problem of imperfect iris recognition, the following issues have been studied in this paper. Firstly, the main steps of iris recognition are introduced, and then the relation between number of features and recognition rate is studied. By designing lots of experiments, iris classification correct rates are studied under the condition of different iris region´s sizes. Results show that different parts of iris region all have importance for iris recognition and iris effective region´ size will determine recognition accuracy.