Synthesis of Large Realistic Iris Databases Using Patch-Based Sampling

This paper presents a framework to synthesize large realistic iris databases, providing an alternative to iris database collection. Firstly, iris patch is used as a basic element to characterize visual primitive of iris texture, and patch-based sampling is applied to create an iris prototype. Then, a set of pseudo irises with intra-class variations are derived from the prototype. Qualitative and quantitative studies reveal that synthetic databases are well suited for evaluating iris recognition systems by achieving three goals: the synthetic iris images bear a close resemblance to real iris images in terms of visual appearance; the proposed framework is able to generate databases with large capacity; statistical performance shows that the synthetic iris images hold all the major characteristics of real iris images.