基于web的人脸识别登陆和管理系统设计与实现

发布时间：2018-01-25 03:17

本文关键词： 人脸识别主成分分析 B/S模式 J2EE　出处：《郑州大学》2017年硕士论文　论文类型：学位论文

【摘要】：如今,数字信息飞速发展,各种信息和交易也趋于数字化和电子化。在信息纷杂众多的同时,信息安全的重要性也在慢慢的凸现出来。基于生物特征的识别技术是信息安全技术中的一个重要分支,因其具有唯一性,可靠性、安全性等众多优点而逐渐成为人们关注的焦点。这其中,基于人体面部特征的识别技术在实际中以其方便、友好和便捷的特点而逐渐被人们所接受。故而,本文构建了一套基于web的人脸识别登陆和后台管理系统。论文的主要工作如下:首先,对国内外人脸识别算法进行研究,给出了基于主成分分析法的人脸识别算法。算法总体分为四步:人脸检测定位、人脸图片归一化处理、人脸特征提取和识别决策。其次,通过对系统业务逻辑进行分析,得出需具备的一些功能性需求和系统指标。其中,基本功能模块有:人脸注册、人脸识别登陆与后台管理。在用户注册模块中,用户注册时能够实时采集人脸图像或上传本地图像进行特征脸的空间的构建;在识别登录模块中,基于主成分分析的人脸识别算法比对待识别拍照图片和人脸库的图片从而判定是否获得登陆权限;在后台管理模块中,管理员需具有对其他管理员信息进行创建及管理的权限、可对人脸注册用户信息完成用户名和人脸查找、修改、更新和删除等常用操作。除此之外,为保证系统能够有效运行及可扩展,还需满足有效性、可移植性、便捷性、安全性及时间特性等系统指标。然后结合业务需求,对系统进行设计,选取B/S模式,采用J2EE的三层架构:表现层、中间层与数据交互层。表现层,主要利用JSP、HTML5、CSS、JavaScript、Ajax等技术完成网页的设计与用户的交互,基于Http协议完成与客户端用户直接交互。中间层采用当下流行的具有MVC模式的Struts2框架对用户请求进行拦截响应和能够实现Java类与数据库表之间进行数据库关系映射的Hibernate框架,数据交互层选用对数据进行管理的关系型数据库MySQL。最后,结合人脸识别算法和系统设计对系统进行实现。通过对算法进行Matlab实验验证,得出人脸识别率为80%,误判率为20%。对系统平台进行搭建,能够实现三大功能性指标和系统运行指标。基于J2EE平台开发的web应用:人脸识别登录及后台管理系统,能够有效保护人们的登录信息、有效识别人脸信息,且登录方便、快捷,还可和其他基于J2EE平台开发的应用相结合,构建整套系统应用。
[Abstract]:Nowadays, with the rapid development of digital information, all kinds of information and transactions are becoming digital and electronic. The importance of information security is gradually emerging. Biometric recognition technology is an important branch of information security technology because of its uniqueness and reliability. Security and many other advantages have gradually become the focus of attention. Among them, the recognition technology based on human facial features is gradually accepted by people because of its convenient, friendly and convenient characteristics. This paper constructs a face recognition landing and background management system based on web. The main work of this paper is as follows: firstly, face recognition algorithms at home and abroad are studied. A face recognition algorithm based on principal component analysis (PCA) is presented. The algorithm is divided into four steps: face detection and localization, face image normalization, face feature extraction and recognition decision. Through the analysis of the system business logic, we get some functional requirements and system indicators. Among them, the basic functional modules are: face registration, face recognition landing and background management. In the user registration module. When users register, they can collect face image or upload local image in real time to construct the space of feature face. In the recognition login module, the face recognition algorithm based on principal component analysis (PCA) is better than the recognition pictures and face database images to determine whether or not to obtain login rights; In the background management module, the administrator needs to have the authority to create and manage the other administrator information, can complete the user name and face search and modify the face registration user information. In addition, in order to ensure that the system can run effectively and scalable, it is necessary to meet the requirements of efficiency, portability and convenience. Security and time characteristics and other system indicators. Then combined with business requirements to design the system, select the B / S model, using J2EE three-tier architecture: presentation layer, middle layer and data interaction layer, presentation layer. The main use of JSP HTML 5 #en0# #en0# JavaScript Ajax and other technologies to complete the design of web pages and user interaction. The middle layer adopts the current popular Struts2 framework with MVC mode to intercept and respond to user requests, and implements the Java class and. Hibernate framework for database relational mapping between database tables. The data interaction layer uses the relational database MySQL. finally, the system is implemented with face recognition algorithm and system design. The algorithm is verified by Matlab experiment. The result shows that the face recognition rate is 80 and the misjudgment rate is 20. The system platform is built. The web application based on J2EE platform: face recognition login and background management system can effectively protect people's login information. Face information is recognized effectively, and the login is convenient and fast. It can be combined with other applications based on J2EE platform to construct the whole system application.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52;TP391.41

【参考文献】