Research And Implementation Of Identity Authentication Technology Based On LPVC And Behavior Characteristics

In recent years with the rapid development of Internet,a variety of opportunities have been brought for social development,economic growth and cultural prosperity.Meanwhile,some risks have also been brought for internet safety.Network security,especially identity authentication technology has gradually become a common concern of the world.For computer and network systems,secure identity authentication is the basic premise to ensure the safe operation of computer network system.Behavioral features have attracted much attention in the field of authentication technology because of the characteristics of uniqueness and easy collection.So it is very meaningful to study identity authentication based on behavior features.Aiming at the deficiency of identity authentication scheme that username/password is widely used in Web platform,a dynamic continuous authentication scheme based on static authentication and dynamic authentication is proposed in this paper.When user login the system initially static authentication must pass.User needs enter the correct username/password and verification code to pass the static authentication.Only through static identity authentication can access to system authority and system resources.After user obtains the system authority successfully,the behavior feature dynamic authentication is activated,which combines keystroke behavior feature with mouse behavior feature.Then it begins to continuously monitor the user operation behavior.By the operating behavior to determine the operating user is the genuine user or imposter user.In general this technique can ensure the security of system login and system after login.The main content of this paper is mainly to discuss generation technology of verification code image and the application of behavior feature in the dynamic continuous authentication system(DCA).A verification code image generation technique based on multi-layer is proposed in this paper,which reduces the accuracy of image recognition for verification codes by computer.Data collection software that can capture user behavior data is developed in this paper.In trust model a method for calculating trust score is improved,which increases the accuracy of trust score calculation.When a system detects input device security,only one input device is detected.In order to prevent a situation where an attacker avoids detection by restricting to one input device,a combination of keystroke dynamics and mouse dynamics is used to identity authentication in this paper.Finally,the proposed approach is validated with extensive experiments.The data used in experiments is collected from 53 users by using the data collection software under complete uncontrolled condition.The test results show that the accurate recognition rate reaches 80%.The method can achieve dynamic continuous authentication.Using the computer to identify the verification code image generated by the traditional technology and the technology based on the multi-layer,the error rate of image recognition based on multi-layer verification code is significantly increased,reaching 20%,effectively preventing a large number of malicious registration procedures.