SaTC: CORE: Small: RUI: Improving Performance of Standoff Iris Recognition Systems Using Deep Learning Frameworks

眼睛虹膜是最准确、独特、通用和可靠的生物识别技术之一，可用于验证个人身份。然而，虹膜识别的准确性取决于数据采集的质量，而数据采集的质量会受到视角、遮挡、扩张等因素的负面影响。由于远距离虹膜识别系统比传统系统受到的限制要少得多，因此捕获的虹膜图像可能会偏角、扩张或在其他方面不太理想。该项目解决了这些具有挑战性的问题，并研究了消除其对防区外系统影响的解决方案。该项目从多个角度提供了潜在的好处：在国家层面，旨在通过提高虹膜识别的性能来引领下一代防区外生物识别系统，从而增强美国的国家安全和竞争力。在州一级，它提高了阿肯色州（EPSCoR（刺激竞争性研究既定计划）州）的研究和教育质量，并为多元化和熟练劳动力的发展做出了贡献。在大学层面，它为来自弱势群体的学生提供研究机会，并为他们提供宝贵的职业技能，包括创造力、自信、批判性思维和解决问题的能力。该项目旨在在传统和非传统虹膜识别框架内使用深度学习技术来提高防区外虹膜识别的性能。首先，开发了基于深度学习的正面图像重建框架，以在将这些图像与数据库中的正面图像进行比较之前消除眼睛结构对隔离图像的影响。它将使用非线性失真图和遮挡掩模在传统虹膜识别框架内展开非理想虹膜图像。其次，非传统虹膜识别框架是基于深度学习算法开发的，以利用眼部和眼周结构中的附加生物识别信息来提高防区外系统的性能。该方法还研究了虹膜/眼/眼周生物识别中凝视角度的影响，并结合了不同远距离图像中的生物识别信息。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

The iris of the eye enables one of the most accurate, distinctive, universal, and reliable biometrics for authenticating the identity of a person. However, the accuracy of iris recognition depends on the quality of data acquisition, which is negatively affected by the angle of view, occlusion, dilation, and other factors. Since standoff iris recognition systems are much less constrained than traditional systems, the captured iris images are likely to be off-angle, dilated, and otherwise less than ideal. This project addresses these challenging problems and investigates solutions to eliminate their effects on standoff systems. The project provides potential benefits from several perspectives: At the national level, it aims to enhance the national security and competitiveness of the United States by improving the performance of iris recognition to lead the next generation of standoff biometrics systems. At the state level, it improves the quality of research and education in Arkansas, an EPSCoR (Established Program to Stimulate Competitive Research) state, and contributes to the development of a diverse and skilled workforce. At the university level, it provides research opportunities for students from underrepresented groups and equips them with valuable skills to build their careers including creativity, self-confidence, critical thinking and problem solving.This project aims to improve the performance of standoff iris recognition using deep learning techniques within both traditional and nontraditional iris recognition frameworks. First, a deep learning-based frontal image reconstruction framework is developed to eliminate the effect of the eye structures on standoff images before comparing these images with their frontal images in a database. It will unwrap non-ideal iris images within the traditional iris recognition framework using non-linear distortion maps and occlusion masks. Second, nontraditional iris recognition frameworks are developed based on deep learning algorithms to improve the performance of standoff systems using additional biometric information in ocular and periocular structures. This approach also investigates the effect of the gaze angle in iris/ocular/periocular biometrics and combines the biometric information in different standoff images.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.