基于触屏行为的智能移动设备身份认证  

Identity authentication of intelligent mobile devices based on touch-screen behavior

作　　者：张彭明 张晓梅[1] 胡建鹏[1] ZHANG Pengming;ZHANG Xiaomei;HU Jianpeng(School of Electric and Electronic Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学,电子电气工程学院,上海201620

出　　处：《智能计算机与应用》2021年第8期154-157,共4页Intelligent Computer and Applications

基　　金：国家自然科学基金(61802252)。

摘　　要：在移动设备的身份认证方面,传统密码认证方式易受到窥肩攻击和屏幕解锁痕迹攻击,不足以保证移动设备隐私安全。本文提出一种基于触屏行为的身份认证方案,获取手指与屏幕交互数据,经预处理后提取触屏行为特征向量,用机器学习算法训练后生成认证模型,对用户进行身份认证。与传统密码身份认证不同,该方案以用户触屏行为信息作为认证密钥,持续检测移动设备用户真实性,即使攻击者看到用户操作行为也很难再现,提高了身份认证安全性。经实验验证,基于触屏行为的身份认证实现了94.42%的认证准确率。In the aspect of identity authentication of mobile devices,the traditional password authentication method is vulnerable to shoulder peeping attack and screen unlocking trace attack,which is not enough to ensure the privacy and security of mobile devices.This paper proposes an identity authentication scheme based on touch-screen behavior,which obtains the interactive data between finger and screen,extracts the feature vector of touch-screen behavior after preprocessing,and generates an authentication model after training with machine learning algorithm to authenticate the user.Different from the traditional password authentication,this scheme uses the user's touch-screen behavior information as the authentication key to continuously detect the authenticity of mobile device users.Even if the attacker sees the user's operation behavior,it is difficult to reproduce,which improves the security of identity authentication.The experimental results show that the authentication accuracy based on touch-screen behavior is 94.42%.

关 键 词：隐私安全 身份认证 行为特征 机器学习 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]