Matching Score Level Fusion for Face and Palmprint Recognition System on Spatial Domain

Face and palmprint recognition system are the most popular biometric system. In certain situations, unimodal systems do not enough to fulfill the security system accuracy and performance. Multimodal is the proper way for providing high performance, security, and accuracy. The appropriate fusion technique also improves the accuracy of recognition system. This paper proposes a development of score level fusion in the spatial domain to enhance recognition system accuracy. In this project Local Binary Pattern (LBP) and Principal Component Analysis (PCA) is used for palm print feature extraction. The Principal Component Analysis (PCA) also has been used for face recognition system. The result of the proposed method for face recognition is 80% and palm print is 85%. The palm print recognition superior than face recognition and a difference is 5%. Fusion of face and palm print is merged using weight sum rule. The experimental result shows that the fusion of these two modalities has improved the recognition accuracy to 98% when tested using ORL and PolyU database.