基于多生物识别考试认证模式的理论与应用研究

【作者】 王琰；

【导师】 李建平；

【作者基本信息】 电子科技大学 ， 计算机应用技术， 2009， 硕士

【摘要】 目前社会考试作弊现象屡禁不止,出现了很多高科技手段的违纪替考行为。传统的考试身份验证流程已经不能完全发现和解决替考问题。基于生物特征的身份认证技术近年来发展迅速,将该技术引入到考试身份认证中来可以有效的控制替考行为,维护考试的公平、公正性。常用的生物特征的有指纹、虹膜、掌型和脸型等,单独依靠其中某一种来进行身份验证会因为人体生物特征缺失等原因造成适用性和准确性降低。本文采用基于指纹和虹膜的多生物特征识别进行考试认证,提高系统的准确性。文章首先研究了基于指纹的识别技术,对指纹图像的采集、图像增强、图像二值化和细化处理、指纹的特征提取和匹配作了阐述。接着对虹膜识别技术进行了研究,分析了虹膜图像采集、图像灰度处理、虹膜定位和归一化处理、特征提取和特征匹配的方法。在对传统方法的研究基础上,为提高系统的效率,提出了一些改进的方法,包括进行指纹图像采集质量的评估、采用基于方向模板的指纹图像方向场计算方法和利用几何线段扫描法来进行虹膜内外圆心的定位方法。多模态生物特征识别技术的关键就是从不同的生物特征角度来进行观察,对每一种生物特征检测定位,提取特征、识别或判决,通过在不同层次将各种特征信息进行融合,对多种生物特征信息进行融合,充分利用各个生物特征提供的信息,从而提高整个识别系统的性能。本文研究了指纹和虹膜这两种生物特征融合,提出了基于隶属度函数的融合算法。文章最后设计了基于多生物识别考试认证的模型,确立了业务流程、功能模块和数据库设计,并对重要的功能模块和数据表及开发方法进行了介绍。更多还原

【Abstract】 At present, the social phenomena of cheating in the examinations have not been prohibited,and there are many high-tech means of committing offence.Traditional process of exam authentication can not fully identify and solve these problems.Authentication technology based on biometrics develops rapidly in recent years.We can effectively control these cheating behaviors by introducing this technology in order to maintain the fairness of the examination.Commonly used biometrics are fingerprint, iris, palm-type and shape of face and so on.If we only depend on one of these means to authenticate,it will reduce the applicability and accuracy because of lacking of human biological characteristics and other reasons.In this paper,the author uses multi-biometric characteristics such as fingerprint and iris to certificate the examination and improve the accuracy of the system.In this paper ,the author studies identification technology of fingerprint and iris. The author puts forward some suggestions in order to improve the efficiency based on traditional methods, which includes the assessment of the quality of fingerprint image acquisition, calculation method by using the template of the fingerprint image based on direction pattern and location method by using scanning method of geometric segment to carry out inner and outer of the circle of iris.The key of multi-biometric identification technology is to observe the biological characteristics from different points of view.We can enhance the performance of the entire identification system in several aspects such as detecting every biological feature, extracting feature, identifying or deciding.In this paper,the author integrates fingerprint and iris, and comes up with fusion algorithm based on membership function.The author finally designs authentication model based on multi-biometrics, establishes a business process,functional modules and database design, and introduces important function modules and data tables, and development methods.更多还原