Abstract:

In order to guarantee the information security of mobile phones, a user authentication scheme based on a mobile phone with threeaxial embedded accelerometer is proposed. It utilizes an accelerometer to capture the acceleration signal of a person’s authenticating gesture.We propose a new algorithm called Difference BottomUp (DBU) to achieve the automatic detection of gesture endpoints. After denoising the acceleration data streams by wavelet packet decomposition, we adopt the Dynamic Time Warping (DTW) method to calculate the similarities between test gesture and reference gesture, and the results of authentication are obtained from these similarities.Compared with the existing methods, the proposed DBU algorithm is more precise in automatic gesture endpoint detection.The authentication experimental results demonstrate that the false rejection rate is below 5% while the false acceptance rate is 0%, which shows the effectiveness of the scheme.

Key words: accelerometer;difference bottom-up(DBU);mobile phone user authentication;dynamic time warping(DTW);biometric