人脸识别技术在期货软件登录中的应用研究

发布时间：2019-03-11 18:22

【摘要】：网上证券期货交易近年来在中国获得了突飞猛进的发展,它具有方便、及时和可获得全面信息等优点,深受广大股民喜爱。然而,个别投资者的账号被不法分子窃取并通过网上交易方式进行盗买盗卖的案例也已屡见不鲜。所以仅仅使用证件、密码口令等传统的方法来进行个人身份鉴别已远远不够,我们亟需更加快捷、更加安全可信的个人身份鉴别方式。本文提出了在期货软件登录界面加入人脸识别的个人身份鉴别方式。此种技术采用了人体本身所特有的生物特征,因此具有更加可靠的安全性和有效性。在信息安全领域,人脸识别技术受到广泛的重视,一些成熟的产品也已陆续投入生产使用。本文首先介绍了多种生物特征识别的技术,通过对比阐述了使用人脸识别技术的原因,简单概括了人脸识别技术目前在某些领域的应用。然后研究了人脸检测的各种方法,并进行了归纳和分析。经过比较后,本文提出并设计了一个实用的由基于肤色分割的人脸粗检测、基于Adaboost的人脸细检测和基于PCA的人脸识别方法组成的人脸识别框架。论文首先研究实现了基于肤色分割的人脸粗检测,然后实现了Adaboost人脸细检测算法,通过创建并训练正负样本,来对人脸样本图像进行训练,得到训练的分类器并对人脸进行检测。在人脸识别部分,本文研究了基于Hu矩和基于PCA的人脸识别方法。尝试了将基于Hu矩的方法引入到人脸识别领域中,最终通过实验比较使用了PCA方法进行人脸识别。最后,将论文中提出的人脸识别框架首次应用到期货软件安全登录中。本论文是基于Intel开发的机器视觉库OpenCV,并采用VC6.0集成开发环境,基于MFC在PC上实现了对系统的开发。实验结果基本达到预想的效果,实现了在期货软件登录系统中加入人脸识别技术的目标。
[Abstract]:In recent years, the online securities and futures trading has been developed by leaps and bounds in China, it is convenient, timely and can obtain comprehensive information and other advantages, it is popular with the majority of shareholders. However, the accounts of individual investors have been stolen by illegal elements and through online transactions to buy and sell cases have also been common. Therefore, it is far from enough to use the traditional methods such as certificate, password and so on to carry out personal identity authentication. We need a faster, more secure and credible way of personal identity authentication. This paper proposes a method of personal identification by adding face recognition to the login interface of futures software. This technology is more reliable and safe because of its unique biological characteristics. In the field of information security, face recognition technology has been paid more and more attention, and some mature products have been put into production one after another. This paper first introduces a variety of biometric recognition techniques, by comparing the reasons for the use of face recognition technology, briefly summarizes the current application of face recognition technology in some fields. Then, various methods of face detection are studied, and summarized and analyzed. After comparison, this paper proposes and designs a practical face recognition framework, which consists of coarse face detection based on skin color segmentation, fine face detection based on Adaboost and face recognition method based on PCA. Firstly, the face coarse detection based on skin color segmentation is studied and implemented, and then the Adaboost fine face detection algorithm is implemented. By creating and training positive and negative samples, the face sample image is trained. The trained classifier is obtained and face detection is carried out. In the part of face recognition, the methods of face recognition based on Hu moment and PCA are studied. This paper attempts to introduce the method based on Hu moments into the field of face recognition. Finally, the PCA method is used for face recognition through experimental comparison. Finally, the face recognition framework proposed in this paper is applied to secure login of futures software for the first time. This paper is based on Intel development of machine vision library OpenCV, and VC6.0 integrated development environment, based on MFC on PC to achieve the development of the system. The experimental results basically achieved the desired results and achieved the goal of adding face recognition technology to the futures software login system.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP391.41

【相似文献】