摘要
对网络用户离线签名身份准确验证,可以保证网络用户信息的安全性,避免伪造签名带来的安全风险。准确验证网络用户离线签名身份,需要提取离线签名图像的特征,通过不同方向对离线签名图像的不同特征隶属函数进行计算分析,达到网络用户离线签名身份准确认证的目的。传统方法在验证用户身份时,未考虑特征隶属函数,导致存在验证效率低、验证准确率低的问题。提出一种新型网络用户离线签名身份验证方法。通过水平方向重心、垂直方向重心、离线签名图像高宽比、签名笔画面积与用户离线签名图像总面积的比、正倾斜度五个特征隶属函数,对用户离线签名图像的特征进行提取,并根据提取结果通过用户离线签名特征向量之间的欧几里得距离对签名相关性的强弱进行衡量,以相关性强弱为依据对用户离线签名样本进行聚类处理,通过最小生成树算法完成用户离线签名的身份验证。仿真结果表明,所提方法提取的特征点多、误拒率和误纳率低,验证所提方法的验证效率高、验证准确率高。
This article puts forward a method of identity verification for offline user signature was proposed based on pattern recognition. The features of user off - line signature image were extracted by five characteristic membership functions such as the horizontal barycenter, the vertical barycenter, the height to width ratio of off - line signature image, the ratio of the stroke area of signature and the total area of user off - line signature image, and the positive obliquity. According to the extraction result, the strength of the signature correlation was measured through Euclidean distance between feature vectors of user off - line signature. Finally, the sample of user off - line signature was clustered based on the strength of correlation. Thus, the identity verification of user offline signature was completed by the minimum spanning tree algorithm. From simulation results, we can see that the proposed method extracts many feature points and has low false rejection rate and false acceptance rate, which proves that the proposed method has high efficiency and accuracy.
作者
孙秋凤
朱婷婷
SUN Qiu - feng;ZHU Ting - ting(Taizhou College,Nanjing Normal University,Taizhou Jiangsu 225300,China;College of Medicine Information,Xuzhou Medical University,Xuzhou Jiangsu 221004,China)
出处
《计算机仿真》
北大核心
2018年第8期378-382,共5页
Computer Simulation
基金
江苏省科学技术厅项目基金(2017300463)
关键词
模式识别
用户离线签名
身份验证
欧几里得距离
Pattern identification
User offline signature
Identity verification
Euclidean distance
作者简介
孙秋凤（1979-），女（汉族）,江苏泰州人，硕士研究生，讲师，主要研究领域为人工智能，模式识别，机器人控制。;朱婷婷（1980-），女（汉族），江苏徐州人，硕士研究生，副教授，主要研究方向：人工智能。
