手指静脉特征识别技术因其具有高精度、高安全性、活体识别和内部特征的优势受到越来越多的关注。然而，在实际应用中，手指静脉识别技术面临着严峻的挑战。一方面，实际应用中的多种因素会降低识别系统的识别精度；另一方面，手指静脉特征能够被成功伪造，使得系统安全性受到威胁。这一现状成为制约手指静脉识别技术推广应用的主要瓶颈。本项目从提高手指静脉识别系统实用性及鲁棒性出发，结合图像处理和机器学习的理论与方法，研究实际应用环境下手指静脉图像质量评估、模板更新和保护的深入理论与实现方法。具体内容包括手指静脉增强图像的质量评估，手指静脉二值图像的质量评估，基于融合理论的质量评估，基于图像质量评估的手指静脉识别，并对注册模板进行更新和保护，建立手指静脉图像质量评估、模板更新和保护的完整理论框架。项目的成功实施，将对手指静脉识别系统的性能改善及实际的推广应用产生积极影响，具有重要的理论意义和工程应用价值。

Due to the advantages of accuracy, security, liveness recognition as well as inner feature, finger vein recognition has been intensively studied. However, finger vein recognition has a serious challenge in practical application: first, many factors can degrade the finger vein recognition accuracy; second, finger vein feature can be faked that affects the system’s security. Both aforementioned problems limit promotion and application of finger-vein recognition. To this end, based on the theory and method in image processing and machine learning, we propose a project to improve practicality and robustness of finger-vein recognition system. It consists of finger vein image quality assessment, template update and protection. Specifically, the main research targets inlude the quality assessment method of enhanced finger vein image, finger vein binary image quality assessment method, the fusion theory based quality assessment method, image quality assessment based finger vein recognition, template update and protection, as well as the presentation of a theoretical finger vein system framework. The successful implemention of project will enhance the performance, promotion and application of finger vein recognition. Therefore, this research is important both from theoretical and practical perspectives.