An Identity Authentication Method Based On Dynamic Gesture And Its Implementation On Smart Phone

With the widespread use of smart phones, the functions and services of phones are getting more and more abundant, and this brings higher security demand of smart phones. Biometric authentication is one of the hotspots of the current research, it has great important application value.Based on the analysis of the current biometric authentication methods and the problems of their applications on smart phones, we propose an identity authentication method based on dynamic gesture, and implement it on the smart phone. The main work and innovations are as follows:1. Considering the identity authentication on smart phones, we propose an identity authentication method based on dynamic gesture. Users complete the authentication through the matches between the gestures to be certified and the gesture template, the authentication process is not affected by the external environment such as light, and it is low cost, easy to accept, safe and reliable.2. We propose a dynamic gesture matching algorithm named EI-DTW, which combines relaxation of the endpoints restriction and early termination of authentication based on Dynamic Time Warping algorithm. EI-DTW algorithm enhances the authentication accuracy by cancelling the endpoint alignment restriction on gesture matching, improves the authentication efficiency by the early termination strategy and limiting the warp slope.3. We propose a design for the prototype system of the identity authentication method based on dynamic gesture. Based on the architecture of classical biometric authentication system, our solution improves the authentication efficiency by adopting the optimal gesture template strategy, design a feature matching method which is based on EI-DTW algorithm and uses a random gesture template. Moreover, our solution automatically updates users’optimal template by their passed gesture, so it is able to adapt the changes of users’gestures in different periods.4. We implement the prototype system on the smart phone with Android platform. Also, we collect more than3000gestures performed by15persons to do experiments. According to the results of the experiments, our authentication method performs very well in authentication accuracy and efficiency, its false rejection rate (FRR) is2.1%, false acceptance rate (FAR) is0, and the average time for the authentication process takes about10ms. Therefore, our method has great practical and promotional value.