Research On Iris Location Algorithms In Mobile Iris Recognition

With the development of many years,iris recognition technology has become mature,and iris recognition system on smart mobile terminal is still a hot research topic.With the rapid development of microelectronics technology,the current smart mobile devices can satisfy the computing power of iris recognition system,but there is still a lack of efficient image recognition algorithm.In the iris recognition system,iris location has always been the focus of research.Therefore,this topic mainly focuses on the iris location algorithm of smart mobile devices.Iris localization is an important step in iris recognition system.The inner and outer edges of iris are obtained by image processing.The inaccurate location will cause a large number of interference from non-iris regions,which will lead to the failure of iris matching process.Therefore,improving the accuracy of iris location is an important research content to improve the performance of recognition system.Most of the methods use iris structural features for localization and texture features for iris matching.However,mobile terminal iris recognition system in visible light has weak constraints on image acquisition,and the iris structure characteristics are vulnerable to various interference factors caused by natural light,resulting in inaccurate positioning.In order to solve the above problems,an iris localization method based on the evolution of regional curve of hybrid geodesic is proposed in this thesis.In the process of image preprocessing,the Lagrange method is used to remove the spot and repair the image.Then,a HOG feature and cascade classifier iris detection model is established to detect the iris region in the whole image,locate the iris initially.Then the linear gray transformation method is used to enhance the image contrast.Finally,the fitting circle method is used to locate the inner edge of the iris,and the mixed geodesic method is used to locate the outer edge of the iris,and a quantitative and objective method to evaluate the accuracy of the iris edge is proposed.In order to simplify the operation process of the algorithm and improve the efficiency of the iris recognition system at the mobile terminal,an improved GrowCut based iris location method is proposed to meet the needs of the iris recognition at the mobile terminal.In this thesis,the algorithm is tested by using the iPhone 5 Library in MICHE-I library.The library is collected by the iPhone 5 mobile phone in visible light.The acquisition process does not impose any constraints on the collected objects.The library contains all kinds of interference encountered in the practical application of mobile iris recognition system,which can test the applicability of positioning methods comprehensively.In this thesis,754 images in the library are used to test the algorithm.The experimental results show that the iris location method based on hybrid geodesic evolution can overcome the influence of natural light on the iris image and locate the iris outer edge accurately.The accuracy of iris location reaches 93.1%.The improved GrowCut method optimizes the label map of segmentation target contour,and the accuracy of localization reaches 92%.Compared with the hybrid geodesic evolution method,the running time of the improved method is reduced by 68%(front lens)and 64%(rear lens)respectively,and is more suitable for the images captured by smart mobile devices.