Iris Recognition: Robust Processing, Synthesis, Performance Evaluation and Applications

The popularity of iris biometric has grown considerably over the past few years. It has resulted in the development of a large number of new iris processing and encoding algorithms. In this dissertation, we will discuss the following aspects of the iris recognition problem: iris image acquisition, iris quality, iris segmentation, iris encoding, performance enhancement and two novel applications.;The specific claimed novelties of this dissertation include: (1) a method to generate a large scale realistic database of iris images; (2) a crosspectral iris matching method for comparison of images in color range against images in Near-Infrared (NIR) range; (3) a method to evaluate iris image and video quality; (4) a robust quality-based iris segmentation method; (5) several approaches to enhance recognition performance and security of traditional iris encoding techniques; (6) a method to increase iris capture volume for acquisition of iris on the move from a distance and (7) a method to improve performance of biometric systems due to available soft data in the form of links and connections in a relevant social network.