A multiresolution approach to iris synthesis

摘要: Databases of human iris images are created and distributed for the purposes testing identification algorithms. For logistical privacy reasons, these databases often too small to fulfill their potential applications.In this work we develop a novel multiresolution approach augment image databases. First, using obtained from reverse subdivision decompose example into set lower resolution components. The components complete representation original consist low approximation characteristics. To generate synthetic combine chosen images. ensure unique, yet realistic, each component is different We quantitatively validate our by employing classical recognition algorithm compare with those that were used create them. results demonstrate effective at augmenting exhibit both visually statistically realistic Graphical AbstractDisplay Omitted Research highlights?A database augmentation.?Realistic visual based on extracted real irises.?Quantitative analysis have similarity irises.?Effectively augments existing new, unique