Abstract
Smartphones have become very popular in recent years. Nowadays, almost everyone has a smartphone. Smartphones are getting more powerful and have more functions than before. You can transfer money, save photos, or buy an application for a smart phone. For these reasons, smartphones have more and more privacy issues than before. Google already proposed a few locking features in an android device, like graphical pattern lock or PIN code lock. Although these kinds of unlock phone features keep the devices safe when the owner is not near the mobile phone, users have to face an inconvenient problem. Whenever an owner wants to use his mobile phone, he has to draw the same pattern or input the same PIN code again and again. It increases the risk of being compromised by an insider. Some previous work has discussed about the implicit authentication. They proposed a new framework that can unlock mobile phones without the owners' effort. In this thesis, we propose a new framework that combines the accelerometer and touch event sensor in a mobile phone to identify the user's identity continuously. All the owner has to do is enroll their behavior biometrics into the mobile device.