基于鼠标键盘行为的身份认证  

Identity Authentication based on Mouse and Keyboard Actions

作　　者：王鑫[1,2] 王梦莹[2] 梁金[2] 

机构地区：[1]桂林电子科技大学海洋信息工程学院 [2]桂林电子科技大学计算机科学与工程学院

出　　处：《中国教育技术装备》2014年第22期50-53,共4页China Educational Technology & Equipment

基　　金：国家级大学生创新创业训练计划项目资助(201210595004)

摘　　要：安全的身份认证是保证计算机及网络系统安全的基本前提,通过研究计算机输入行为特征,提出并实现基于鼠标键盘行为的身份认证系统。该系统运用优化的SVM算法分析用户信息,结合鼠标和键盘双指标复合认证。系统设计了友好的训练场景,明显缩短了训练和认证的行为数据采集时间。实验证明,双指标复合认证与单指标相比较而言,准确性有了较大提高,系统误识率和拒识率都有所降低。The security of identity authentication is the basic premise of the safety of computer and network system. This paper proposed and realized the identity authentication system based on mouse and keyboard actions by studying the trait of behavior, using the optimized SVM algorithm to analysis user information, adapting the mouse and keyboard double indicator compound authentication. The system shortened the time of training and acquisition certiifcation of behavior data greatly by designing a friendly training scene. Experiments proved that the double indicator compound authentication accuracy had a larger increase compared with single index, and decreased the system course rate and false reject rate.

关 键 词：支持向量机 行为特征 身份认证 双指标 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]