指静脉识别算法及其密码应用

发布时间：2018-05-27 09:26

本文选题：生物特征识别 + 手指静脉识别　；参考：《杭州电子科技大学》2017年硕士论文

【摘要】：近年来,随着信息技术的快速发展,信息安全受到人们越来越多的关注。生物特征识别技术作为保障信息安全的一种重要途径,在信息安全领域具有不可替代的地位。与传统的基于指纹、虹膜、人脸识别等生物特征识别技术相比,手指静脉识别技术作为一种新的生物特征识别技术,主要特点在于将手指上隐藏的静脉血管特征作为识别的特征。除此以外,手指静脉识别还具有以下三个特点:活体识别、高安全性和非接触性。由于以上种种优势,手指静脉识别已经成为一个新的研究热点,因此指静脉识别的相关研究具有重要的实际意义。本文首先介绍了指静脉识别系统中的基本内容和一些经典算法;其次,考虑到指静脉系统在采集图像时手指的旋转问题与系统安全性问题,本文提出了一种基于Radon变换和混沌加密的指静脉识别算法;此外,本文研究了模糊金库算法,并提出了一种基于指静脉特征的模糊金库方案。具体工作如下:1.研究了指静脉识别算法的主要内容,主要包括预处理、指静脉图像增强算法、纹路分割算法和生物特征识别算法中常用的性能分析指标。2.提出了一种基于Radon变换和混沌加密的指静脉识别算法。该算法首先对静脉图像进行图像增强、Niblack分割、填充、细化一系列操作得到指静脉骨架图像,然后根据Radon变换计算得到该指静脉的基本方向,根据该基本方向对指静脉骨架进行旋转校正。在注册时,利用用户输入的密钥生成Logistic混沌序列对指静脉骨架特征进行加密,并将加密后的特征图像保存于特征数据库中。在匹配时,用户需采集待验证指静脉图像并输入密钥,对采集的指静脉经过处理得到特征信息,再利用密钥解密保存在数据库中的指静脉特征,计算两特征矩阵间的Pearson相关系数进行匹配。实验结果表明:该算法在不经过大量模板训练的情况下也能取得较高的识别精度,且该系统的安全性更高。3.提出了一种适用于指静脉的模糊金库算法。该算法分为两个阶段:密钥绑定和密钥恢复。在密钥绑定时,使用细节点生成方法得到指静脉细节点,并定义了新的参数“长度”,使用(x,y,l)描述每个细节点,再添加大量的杂凑点对真实点进行隐藏,对每个点(包括真实点和杂凑点)进行一系列数据处理和转换后存入模糊金库中,其中对真实点利用密钥绑定多项式,而对杂凑点则利用随机数计算。在恢复密钥时,通过比对查询指静脉中的细节点和模糊金库中的点的坐标信息,选择相匹配的细节点并进行相反计算,求得多项式并恢复出保护的密钥。通过实验和分析表明:该模糊金库算法具有良好的实用性和安全性。
[Abstract]:In recent years, with the rapid development of information technology, people pay more and more attention to information security. As an important way to guarantee information security, biometric identification plays an irreplaceable role in the field of information security. Compared with traditional biometrics such as fingerprint, iris and face recognition, finger vein recognition is a new biometric technology. The main feature is that the hidden veins on the finger as the recognition feature. In addition, finger vein recognition has three characteristics: in vivo recognition, high safety and non-contact. Because of the above advantages, finger vein recognition has become a new research hotspot, so the research of finger vein recognition has important practical significance. This paper first introduces the basic contents and some classical algorithms of digital vein recognition system. Secondly, considering the problem of finger rotation and system security when the digital vein system collects images, In this paper, a digital vein recognition algorithm based on Radon transform and chaotic encryption is proposed, in addition, the fuzzy vault algorithm is studied, and a fuzzy vault scheme based on finger vein features is proposed. The work is as follows: 1. In this paper, the main contents of digital vein recognition algorithm are studied, including preprocessing, enhancement algorithm of digital vein image, texture segmentation algorithm and biometric recognition algorithm. A digital vein recognition algorithm based on Radon transform and chaotic encryption is proposed. In this algorithm, the vein image is segmented, filled, and a series of operations are thinned to get the skeleton image of the digital vein. Then the basic direction of the digital vein is calculated according to the Radon transform. The digital venous skeleton was rotated and corrected according to the basic direction. At the time of registration, the Logistic chaotic sequence is generated by the user's key to encrypt the digital vein skeleton feature, and the encrypted feature image is stored in the feature database. When matching, the user needs to collect the image of the digital vein to be verified and input the key. After processing the collected digital vein, the characteristic information is obtained, and then the digital vein feature stored in the database is decrypted by using the key. The Pearson correlation coefficient between the two characteristic matrices is calculated to match. The experimental results show that the algorithm can achieve high recognition accuracy without a lot of template training, and the security of the system is higher. 3. A fuzzy vault algorithm for digital veins is proposed. The algorithm is divided into two stages: key binding and key recovery. In key binding, we use the detail generation method to get the minutiae of the finger vein, define the new parameter "length", describe each detail point by using a number of hash points, and hide the real point by adding a large number of hash points. Each point (including real point and hash point) is processed and converted and stored in the fuzzy vault, in which the key binding polynomial is used for the real point and the random number is used for the hash point. When the key is restored, by comparing the coordinate information between the point in the finger vein and the point in the fuzzy vault, the matching details are selected and the reverse calculation is carried out, and the polynomial is obtained and the protected key is restored. Experiments and analysis show that the fuzzy vault algorithm has good practicability and security.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP309

【参考文献】