一个基于HMM的动态手写签名认证系统的设计与实现  

THE DESIGN AND IMPLEMENTATION OF A DYNAMIC HANDWRITTEN SIGNATURE VERIFICATION SYSTEM BASED ON HMM

作　　者：袁余良[1] 沈峰[1] 杨飞 潘金贵[2,1] 

机构地区：[1]南京大学计算机科学与技术系,南京210093 [2]南京大学软件新技术国家重点实验室,南京210093

出　　处：《模式识别与人工智能》2004年第2期146-150,共5页Pattern Recognition and Artificial Intelligence

摘　　要：动态手写签名认证技术作为一种有效的生物认证技术,具有不易遗忘、模仿困难和数据获取简单自然等优点,本文介绍了一个基于隐马尔可夫模型(HMM)的动态手写签名认证系统的设计与实现,该系统将签名认证和识别相结合,简化了认证过程.同时能自动学习,适应用户书写习惯的变化,从实验数据来看,其性能是较满意的.Dynamic handwritten signature verification, as an effective and efficient kind of biometric based authentication schemes, has the advantages of being unforgettable, difficult imitated and natural to collect data. This paper introduces the design and implementation of a dynamic handwritten signature verification system, which is based on hidden markov modal (HMM). The system combines verification with recognition, simplifies the process of verification. It has the function of automatic learning and can adapt to the change of writers' style. The performance is satisfying with the data of experiment.

关 键 词：隐马尔可夫模型 生物认证 动态手写签名认证 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]