Browsing by Author "Hasimah Ali"

Single biometric verification systems often have to contend with spoof attacks, noisy data, non–universality of the biometric trait and unacceptable error rates. Any attempt to improve the performance of such system in many situations may be ineffective due to various problems. The objective of this thesis is to design and develop of an integrated biometric system for high level security as well as analyze of the overall systems performance. An integrated biometric system would provide anti–spoofing measures by making it difficult for an intruder to spoof multiple biometric traits simultaneously. Therefore, the proposed integrated biometric system is a combination of keystroke pressure–based typing biometric with iris recognition system for high level security purpose. The proposed integrated biometric system is designed to operate in parallel mode and the integration process is made at decision level. An Artificial Neural Network (ANN), Adaptive Neuro–Fuzzy Inference System (ANFIS) and Support Vector Machines (SVMs) are adopted as classifier to verify the authorized and unauthorized users based on extracted features of individual traits. Combining multiple classifiers is implemented for integrated biometric systems to improve the system accuracy as well as enhanced the robustness of the verification system. The experimental results gave 0% FRR (False Rejection Rate) and 0% FAR (False Acceptance Rate). This indicate that the integrated biometric system improve the overall performance of the verification system.