Research On User Authentication Based On The Features Of Unlocking Gestures On The Touchscreen

With the development of science and technology,many business activities are proceeding through the internet on mobile devices.These activities involve much private information of the users.The trends in using smart mobile devices with touchscreen are putting private information of the users at high risk,especially when the mobile device is hacked or stolen.Because users are aware of the importance of private information on mobile devices,many users have put much attention to the authentication schemes for mobile devices.Among the methods of user authentication for mobile devices with touchscreen,because of the simplicity and convenience,gesture authentication is one of the most common methods.However,the gesture authentication loses its safety when the gesture used for authentication is leaked to an invader.Therefore,how to improve the security of gesture authentication is an important issue.Concerning the security risk of leaking of the unlocking gesture,a user authentication methodology based on the features of unlocking gesture on the touchscreen is proposed.The research in this paper is organized as follows:(1)A user authentication methodology based on the features of unlocking gesture on the touchscreen is proposed,and these features contain time(including the dot hold time,the line time and the total time),pressure,touch size and sliding speed features.(2)In this paper,we constructed verification and identification experiments,and2 types of gestures are used by our experiments.In the verification experiment,a statistic-based threshold training methodology is proposed to authenticate a test sample belonging to the original user or the invader.In the identification experiment,a kNN(K Nearest Neighbor)method based methodology is proposed to classify the test sample into a certain user.(3)In the verification experiment and the identification experiment,we get ideal experiment results.Then in this paper,a coefficient of variation analysis of each feature and the analysis and discussion of the experiment results are performed.Finally,the conclusion and expectation of this paper are proposed.