A Smart phone Identification Method Based on Gesture

School of Electronic Science and Engineering, National University of Defense Technology, China

Abstract

In order to promote the practicality of smart phone identification based on gesture, we introduce weighted morphological characteristics and early termination of authentication into dynamic time warping (DTW) algorithm, which based on characteristic moving data captured by in-built accelerometer and gyroscopes, and put forward an effective identification algorithm (ME-DTW) for smartphone. The algorithm controls the contribution made by the difference between the latitudes to the total Euclidean distance by smartphone morphological characteristics. It also introduces restricted areas touch trigger gesture acquisition scheme and authentication gesture length selection scheme based on normal distribution, which effectively improve the identification accuracy and efficiency. Experimental results show that: when others imitate gestures attack false acceptance rate tends to 0%, the personal identification false rejection rate remained at 3.29%, which can meet most practical security needs