Design Implementation Of Continuous Identity Authentication System Based On Multi-mode Recognition

With the rapid development of computer technology and the traction guidance of the "Internet+" strategy,Internet technology has gradually entered various areas of the traditional industry.At the same time,the data security and information privacy issues of the traditional Internet have gradually expanded to all aspects of society.As the first line of defense,identity authentication technology is very important for information security.Typical identity authentication technologies include password-based password authentication,ID card-based holder authentication,fingerprint-iris-based physiological characteristic authentication,and gait-signature-based behavior characteristic authentication.The first three authentications are "one-time authentication" technologies,in which the user's identity is verified only when the user logs in to the system for the first time.Once the authentication is passed,the user's operation is no longer blocked.Therefore,when an illegal user completes an authentication through identity forgery,it will continue to use system resources without any obstruction.The biometric authentication technology based on behavior characteristics will continuously collect behavior characteristics during the user's use,and verify the user's identity,and ensure that the user's identity in the process of using the device is legal,and prevent the identity fraud of illegal users to avoid data privacy leakage and reduce information security risks.Therefore,this paper presents a continuous identity authentication system to continuously monitor user operations since the device is started.Once the system verifies that the user's identity is abnormal,the device will be automatically locked and issue an alarm to achieve effective prevention of "one-time authentication"risk and upgrade guarantee for user information security.In terms of system design,this paper has fully investigated the research progress at home and abroad.Considering that the technical solution of single modal recognition is not enough to effectively describe the user's behavior habits,there is a large room for improvement in accuracy and recognition rate.Biometric recognition technology of keystroke dynamics and mouse dynamics and application recognition technology based on user activity records serve as system technical support.Multi-modal fusion of three types of identification technologies is introduced,and a trust model is introduced to generate a continuous identity for sustainable authentication.system.The main research content of this paper is the design and implementation of continuous identity authentication system based on multi-mode recognition.Main work:(1)Analysis of the functional and non-functional requirements of the multi-mode authentication system based on the application scenario;(2)Overall design and hierarchical design based on the implementation of the multi-mode authentication system,dividing the system into data The four layers of processing layer,data transmission layer,algorithm analysis layer and application layer;(3)A complete system implementation based on the layered design,and the keystroke model,mouse from the whole process of data collection,feature selection and model training The detailed implementation of key functional modules such as models is explained in detail;(4)Multi-mode identity authentication system is fully tested.Finally,this paper summarizes the shortcomings of the system and the next improvements.