Abstract:
This paper proposes and implements a multilayer authentication and template protection scheme known as the hybrid scheme. The hybrid scheme is based on the integration of modified shielding function with matrix transformation. It is a 2-in-1 approach which provides good recognition accuracy, security of stored templates and user privacy. The recognition accuracy of the proposed approach is evaluated using a multi-biometric dataset obtained from the fusion of face and iris images. The multibiometric dataset is constructed
using feature level fusion to combine face images in CASIA Near Infrared Database and iris images in CASIA Iris Database version 1. Security analysis is used to determine the robustness of the hybrid scheme against security threats such as guessing and key exhaustion (or brute force) attacks. This is carried out using complexity analysis to determine the level of difficulty an attacker faces in an attempt to launch compromise the authentication scheme.
The privacy capability of the hybrid scheme is analyzed using the requirements proposed in the ISO/IEC JTC 1/SC 27 standard. Experimental results as well as security and privacy analyses show that the approach provides good recognition accuracy as well as high-level template security and user privacy.
Keywords: Authentication, Biometric, Privacy, Recognition Accuracy, Security.
