Dynamic iris localisation: a novel approach suitable for fake iris detection

Abstract

In an iris recognition system, accurate iris segmentation and localisation from an eye image is the foremost important step. Success rate of any feature extraction algorithm of an iris recognition systems is primarily decides by the performance of iris segmentation from eye image. In this paper a robust and efficient method of iris segmentation is proposed. In the proposed method, the outer boundary of an iris is calculated by tracing objects of various shape and structure. For inner iris boundary, two eye images of same subject at different intensities are compared with each other to detect the variation in pupil size. Based on the pupil size variation, the inner boundary of iris is detected. The variation in pupil size is also used for aliveness detection of iris. Thus, this approach is a very promising technique in making iris recognition systems more robust against fake-iris-based spoofing attempts. Outer iris boundary detection method also detects intersection of iris with upper and lower eyelids to ensure minimum loss of iris information. The algorithm is tested on Phoenix database of 384 images both eyes of 64 subjects . The success rate of accurate iris localisation from eye image is 99.48% with minimal loss of iris texture features in spatial domain as compared to all existing techniques. The processing time required is also comparable with existing techniques. (4 pages)