Design And Realization Of Access Control System Based On Face Recognition Technology

Face recognition is a hot subject in the study fields of security technology. This article, integrating face recognition technology into access control system, brings forward a new scheme which can promote the security of system.The new system adopts the face recognization technology as its authentication part and RBAC based on identity code and double-way authentication as its permission assignment part.In face detection part, this article brings forward a quick, efficient and steady face detection algorithm based on skin color and geometry characteristics information. First, this algorithm adopts a coarse to fine strategy, the face candidates is detected first by skin color information, then the face candidates are cross-checked by geometry characteristics. This algorithm is proved to be efficient and robust to the images in different background, size and rotation angles.In face recognition part, this article adopts the arithmetic of PCA and BPNN and introduces the concept of reference Eigenface aggregate to enhance the speed of face recognition. The distance vector of face candidate to the reference Eigenface aggregate is calculated firstly, and then a small aggregate of faces can be selected by the vector. At last, the face candidate validates the small aggregate one by one with BPNN trained before. This scheme is very efficient when the number of face is great.In access control part, this article adopts role-based access control (RBAC) and brings forward a new scheme for access control with identity code, role environment function and double-way authentication based on Harn digital signature, which solves the problems of public permission and private permission inherited independently in role hierarchies, dynamic constraints for role's permission and system security. The new scheme is secure and easy to implement.