基于PCA和LDA算法的人脸识别考勤管理系统的设计与实现

发布时间：2018-05-24 02:09

本文选题：PCA算法 + LDA算法　；参考：《湖南大学》2016年硕士论文

【摘要】：伴随着科技水平的进步,管理方式的科技含量越来越高,但能否快捷、准确的获取到员工的考勤信息仍然是很多企业人力资源管理部门所急需解决的问题。而且目前,仍有少数企事业单位依然使用传统的方式对员工的出勤信息进行管理,容易出现"脏数据"的现象,因此建立一套方便、快捷且满足企业实际的考勤管理信息系统逐渐成为企业事业单位考勤管理的一个势在必行的发展方向。本文首先分析了人脸识别技术的研究背景和现状,从系统的整体着眼,结合了企业对考勤管理的业务流程,分析了人脸识别考勤系统的可行性,给出了功能需求和非功能需求,同时还介绍了实现系统的关键技术如Adaboost算法、PCA算法和LDA算法的研究。在系统的设计中,采用了统一建模语言对系统进行了总体的设计,包括系统的逻辑架构、功能架构、网络架构以及数据库的设计,还利用流程图和时序图对系统各个功能模块进行了详细设计。针对PCA算法提取特征时容易发生小样本问题,提出了融合PCA和LDA算法来提取人脸特征,并将该算法运用到人脸识别考勤管理系统中。最后介绍了系统的实现,系统采用客户端/服务器体系结构,实现了对员工考勤的管理。系统分为三个功能模块:考勤管理、基本信息管理和考勤统计管理,三个模块之间相辅相成,实现了员工考勤的信息化管理。系统测试和运行结果显示,系统界面友好、操作简单、数据可靠,不仅能够满足企业考勤管理工作的需求,还优化了企业考勤管理的流程,降低了管理成本,有效地提高了企业的管理效率,同时,严格的数据获取方式与数据检测流程保障了考勤管理系统中数据的完整性与安全性。
[Abstract]:With the progress of science and technology, the scientific and technological content of management mode is higher and higher, but it is still an urgent problem for many enterprise human resources management departments to get the attendance information of employees quickly and accurately. And at present, there are still a few enterprises and institutions that still use the traditional way to manage the attendance information of their employees, which is prone to the phenomenon of "dirty data", so it is convenient to establish a set. Fast and satisfying the actual attendance management information system has gradually become an imperative development direction of attendance management in enterprises and institutions. This paper first analyzes the research background and present situation of face recognition technology, from the whole system, combined with the business process of enterprise attendance management, analyzes the feasibility of face recognition attendance system, and gives the functional and non-functional requirements. At the same time, the key technologies of the system, such as Adaboost algorithm and LDA algorithm, are introduced. In the design of the system, the general design of the system is carried out by using the unified modeling language, including the logical structure, the functional architecture, the network architecture and the database design of the system. In addition, each function module of the system is designed in detail by flow chart and timing diagram. In order to solve the problem of small sample when PCA algorithm is easy to extract features, a fusion of PCA and LDA algorithms is proposed to extract face features, and the algorithm is applied to face recognition attendance management system. Finally, the implementation of the system is introduced. The system adopts the client / server architecture, and realizes the management of employee attendance. The system is divided into three functional modules: attendance management, basic information management and attendance statistics management. The three modules complement each other and realize the information management of employee attendance. The system test and running results show that the system has friendly interface, simple operation and reliable data. It can not only meet the needs of enterprise attendance management, but also optimize the process of enterprise attendance management and reduce the management cost. At the same time, strict data acquisition and data detection flow ensure the integrity and security of data in attendance management system.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.52;TP391.41

【参考文献】