Combination of Multiple Biometrics for Recognition of Persons

Abstract

The security of information is ensuring that the only authorized users are able to ac- cess the required contents, thereby entails con dentiality of exchange information. The recognition of the person identity is one means to ensure this purpose. In fact, due to the great need for such recognition, man has developed several ways that are related to information's that a person has or knows. However, to overcome the limitations associated with such traditional means, other means of security has been developed that allow obtaining the speci c information of the person. It is the biometrics-based recog- nition. Biometric technology has attracted a great attention in recent years. In the biometric security systems, the personal identity recognition depends on their behavioral, biolog- ical or physical characteristics. Currently, a number of biometrics technologies are developed and one of the most popular biometric trait is Finger-Knuckle-Print (FKP) due to the user-friendly and the low cost. This thesis presents a new approach, where the simple deep learning is applied to create a multi-modal biometric system based on images of FKP modalities which extracted their features by Principal Component Anal- ysis Network (PCANet) and Discrete Cosine Transform Network (DCTNet). In the proposed structure, PCA or DCT is employed to learn two-stage of lter banks followed by simple binary hashing and block histograms for clustering at feature vectors, which is adopt as input for classi cation. Thus, the Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) classi ers are used for the PCANet and DCTNet features, respectively. To improve the recognition rates, a multimodal biometric system based on matching score level fusion scheme was generated. Using an available FKP database, we conducted a series of identi cation experiments and the obtained results show that the design of our identi cation system achieves an excellent recognition rate and having a high anti-counterfeiting capability.