基于特征向量和分类树的签名认证方法  

Signature Verification Based on Feature Vectors and Clustering Trees

作　　者：邱德红[1] 陈杲[1] 陈传波[1] 金先级[1] 

机构地区：[1]华中科技大学计算机科学与技术学院,湖北武汉430074

出　　处：《计算机工程与科学》2002年第4期48-49,70,共3页Computer Engineering & Science

摘　　要：本文介绍了如何从签名的基本信息中提取签名特征构成特征向量 ,在此基础上提出了生成分类树 ,以及根据分类树的树型特征进行签名真伪认证的方法。该方法简单快捷 ,决策依据的物理意义明确 ,在实验中获得了较好的认证效果。This paper describes how to extract features from the basic information of signatures. Based on feature vectors, a clustering tree is generated. The shape of the clustering tree is then analyzed to decide whether the signature being verified is genuine or false. This method is simple but effective, and its physical meaning is clear. A good result is obtained in the experiment of signature verification.

关 键 词：特征向量 分类树 签名认证 信息提取 信息处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]