Research On Continuous Identity Authentication Based On User Behavior

Identity authentication is a process to determine whether a specific person or device can access the privacy data stored in the system,application or device,and is the key to ensure the security of intelligent mobile devices.The mechanism of continuous authentication can continuously capture the biometric or behavioral characteristics of users to verify the identity of users during the use of devices,which can eliminate the security risks brought by one-time identity authentication mechanism.In this thesis,based on the characteristics of user behavior,the method of continuous identity authentication is studied.According to the user's behavior data generated when interacting with the touch screen when using the intelligent mobile device,a method of continuous identity authentication based on touch gesture is proposed,which can improve the security of the intelligent mobile device.The main research contents and contributions are as follows:1.A feature extraction method based on touch gesture data is proposed.Combined with the hybrid feature selection scheme of mutual information model and recursive feature elimination model,the optimal user behavior feature is selected to eliminate redundant attributes.2.For the problem that training user identity model needs a large number of data samples,this thesis applies the data enhancement technology based on jitter to the identity recognition method,so that only a small sample of data can be used to train the identity model with high recognition accuracy,and it can save the data collection time when users register the identity model,and improve the user experience.3.Aiming at the security problem of user behavior data,combined with the fuzzy safe algorithm in the biometric template protection method,the integral function is used to transform the polynomial in the fuzzy safe,and a behavior data protection scheme based on the improved fuzzy safe algorithm is proposed.The experimental results show that the FRR and FAR of the proposed scheme are0.57% and 0.98% respectively.