基于击键特征的用户身份认证新方法

A New Approach for User Authentication Based on Biometrics

口令保护机制是绝大多数计算机及网络安全管理中的重要一环。如何有效防止因口令失窃而造成计算机及网络系统的非法进入，一直是一个倍受关注的问题。为此文章提出了一种利用用户口令输入击键特征进行用户身份认证的新方法。该方法利用人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独有特征的击键特征向量，并由此提出了一种基于有限正例样本集合识别正反例的新算法。相关实验结果表明该身份认证新方法具有较高的用户识别能力。

Password plays an important part in security management for most computers and networks. How to prevent illegal intrusion due to password stealing is always one open question. A new approach which identifies user by utilizing keystroke features is proposed when user inputs his password. The key pressure and keystroke frequency when user strikes the keyboard to enter his password are used to construct a peculiar keystroke vector representing each user 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 discerning ability.