Biometric Key Generation Research Based On Dynamic Handwritten Signature

With the rapid prevalence and development of computer network and communication technology, information security is more and more important. Biometrics and information security complements each other. Biometrics means physiology feature or behavior feature, associating with people physically, which is the only way to identify one person sufficiently. At present, Biometrics is a hot spot in information security domain. In this dissertation, key generation based on dynamic signature is researched, combining the newest F-Tablet which can acquire both track and writing force. The main work of this dissertation can be summarized as follows:Firstly, biometrics related is expatiated and present research status of key generation based on biometrics is probed into.Secondly, signature samples are acquired using force tablet F-Tablet and signature database is built. The database is composed of true signature samples, professional false signature samples and simple false signature samples. According to the algorithm, Gauss filter is chosen as the pretreatment scheme.Thirdly, key generation algorithm based on F-Tablet signature database is put forward. According to the writing force, track, velocity and acceleration of dynamic signature acquired by F-Tablet, the stable and meaningful feature is extracted using Fourier transform, and then key is generated using multi-state discretization.At last, performance is analyzed from authentication performance and security performance. Authentication performance is verified from reference sample number, signature complexity and signature variance. Security performance is studied from the exhaustive search, key statistics attack and feature statistics attack. Experiment indicates that the algorithm is feasible and effective.