الفهرس | Only 14 pages are availabe for public view |
Abstract This thesis is concerned with a vital topic in securing communication networks. Our main concern is how to access the networks. Security tools are needed for this task. It can be implemented via a biometric verification system. The utilization of open biometrics in public databases is not recommended. Hence, there is a need to use cancellable biometric systems. This thesis presents an elegant form for generating cancelable biometric templates through including some intended distortion in biometric templates to keep the original ones away from hacking attempts. This objective is achieved with the help of adaptive filtering algorithms and quaternion mathematics. Moreover, optimization tools are also used to obtain the best biometric system performance. Two adaptive filter algorithms are utilized in this thesis to generate the cancellable biometric templates. They are the Least Mean Square (LMS) algorithm, and the Leaky Least Mean Square (LLMS) algorithm. The main idea is to use the biometric template as an input to the adaptive filter with an arbitrary desired response. Hence, the cancelable biometric template is extracted from the weights of the filter as a signature for the original template. All adaptive filters used are built with quaternion mathematics. Whale optimization is used to select the best value for the weights used in the LLMS algorithm. A comparison study is presented between the results of the proposed algorithms and those obtained with an ordinary implementation of quaternion mathematics for cancelable biometrics. Simulation results prove that the cancelable biometric system based on the LLMS algorithm and quaternion mathematics with optimization achieves the best performance. |