User Reauthentication On Smartphones Via Behavioral Biometrics

Context-awareness is a highly integrated research that has been focused more deeply in many research fields such as activity recognition etc. With the emergence of smartphones as an essential part of daily life and the development of computing and sensing ability. Not only can we imply context-awareness on smartphone, it also contains enormous research and the commercial value. Besides, because people save lots of personal messages on their devices(such as smartphone), the demand for user authentication has increased manifolds. The effective and widely practiced biometric schemes are based upon the principle of who you are which utilizes inherent and unique characteristics of the user. In this context, the behavioral biometrics like Sliding Dynamics and Pressure Intensity make use of on-screen sliding movements to infer the user‘s patterns.In this thesis, we present Safeguard, an accurate and efficient smartphone user authentication(verification) system based upon on-screen finger movements. The key feature of proposed system lies in fine-grained on-screen biometric metrics, i.e. sliding dynamics and sliding pressure, which are unique to each user under diverse scenarios. Besides, the behavior biometic metrics is born to every one, we can't change it by our own consciousness. So it's suitable to the authentication process. In the implement, after the data collection, pre-processing, feature selection and feature process, we determine three behavior biometic metrics(Angle, Distance Ratio and Sliding Pressure) which is based on sliding. Then we first implement our scheme through five machine learning approaches and finally select the Support Vector Machine, SVM, based approach due to its high accuracy. We further analyze Safeguard to be robust against adversary imitation attack and calculate the tiny system overhead on computing time and storage. We validate the efficacy of our approach through implementation on off-the-shelf smartphone followed by practical evaluation under different scenarios(such as different applications and devices). The application on Android is practicable on various scenarios. We process a set of more than 50,000 effective samples derived from a raw dataset of over 10,000 slides collected from each of the 60 volunteers over a period of one month. The experimental results show that Safeguard can verify a user with 0.03% false acceptance rate, FAR, and 0.05% false rejection rate, FRR, within 0.3 sec with 10 to 20 slides by the user. Comparing to the European Standard for Access Control Systems(FAR less than 0.001% and FRR less than 1%), our work has made great progress and we will optimize it in the future.