摘要
口令保护机制是计算机系统及其网络安全管理中重要的一环。如何有效防止因口令失窃所造成计算机及其网络系统的非法进入,一直是一个倍受关注的问题。为此本文提出了一种利用用户口令输入时的击键特征进行用户身份认证的方法。该方法利用用户在点击键盘时所产生的按键压力和击键频率,构造出能够描述每个用户特有的击键特征向量,并由此提出了一种基于有限正例样本集合识别正反例的归纳方法。相应的实验结果表明这种用户身份认证方法,具有较高的用户识别能力。
The password plays an important part in security management for computers and networks. How to prevent illegal intrusion due to password stealing is always one open problem. In this paper, a new approach is proposed to identify user by utilizing his keystroke features generated by user password input process. The keystroke pressure and frequency when user strikes the keyboard to enter his password is used to construct a peculiar keystroke vector representing each user identity respectively. A new algorithm is put forward to classify positive and negative examples only by a limited positive examples. The related experiment indicates this approach has a good user discerning ability.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2003年第2期252-256,共5页
Pattern Recognition and Artificial Intelligence
