Towards an automatic gait recognition system using activity recognition (wearable based)

The need of increasing the security measures in mobile devices has led researchers focus on finding new security mechanisms. In this paper we propose a solution to keep data secured by ensuring that only the authorized user can access the data in a mobile phone. By using gait recognition as an important element for the authentication process we propose an automatic gait recognition system to be used for continuous authentication. Since recent gait recognition research only focuses on manual extraction of walking activities from the accelerometer signal, a solution to this issue could be activity recognition that would reduce the disadvantages of gait recognition by identifying the activities of a person continuously and automatically. Activity recognition would not only make it possible to authenticate the user in different daily activities like slow walking, normal walking, fast walking even running, but also help in avoiding authentication when the user is in passive state like sitting, standing still, etc.. This is one of the key factors and an interesting challenge which would benefit the data security area.