四维光场虹膜图像的预处理与分类识别

光场成像技术的快速发展为虹膜识别学科创新、技术进步和产业发展提供了新机遇，但是新数据也带来了新问题——如何表示、处理和识别四维光场虹膜图像。本项目旨在创建比较系统的四维光场虹膜图像预处理与分类识别的理论和方法，研究光场虹膜图像的数据库建设、数据解析、特征表达、预处理、分类识别等内容，充分发掘四维光场数据的“三多”特性（多维度、多视角、多焦点）来解决复杂场景虹膜识别的瓶颈问题（例如拓展景深、虹膜定位、眼皮分割、睫毛遮挡、超分辨率、活体检测、种族/性别分类、身份识别等），建设光场虹膜识别原型系统验证新理论和新方法。本项目有望开辟虹膜识别学科在光场成像时代的新方向，抢占虹膜识别前沿科技和产业的制高点，为国家公共安全、金融、通关、社保等领域提供自主知识产权的虹膜识别新技术。

Development of light field (LF) imaging technology provides new opportunities to iris recognition. But a new problem is how to represent, process and recognize iris images in 4D light field. The project aims to establish systematic theories and methods for iris image processing, classification and recognition. We will construct a large LF iris image database and develop algorithms to analyze, represent, process, classify, and recognize LF iris images. The features of LF data such as multi-dimension, multi-view and multi-focus will be used to solve the challenging problems in practical iris recognition applications, such as extending depth of field, iris localization, eyelids segmentation, super-resolution, liveness detection, race and sex classification and personal identification. A demo system of LF iris recognition will be developed to evaluate the novel algorithms. The research achievements of this project will promote both academic research and industrial development of iris recognition. LF iris recognition technology has wide applications in public security, banking, border-crossing and welfare distribution, etc.