الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 170 |  |  |
| | | from | 170 | | |
Abstract
A wide variety of systems require reliable person recognition schemes to either confirm or determine the identity of an individual. Though biometric systems have been successfully deployed in a number of real-world application, biometrics is not yet a fully solved problem. The main factors that contribute to the complexity of biometric system design are convenience, accuracy , scalability and security. The grand challenge in biometrics is to design a system that operates in the extremes of all these factors. The objective of this work is to design a robust multimodal biometric system to meet the stringent requirements imposed by these challenges.
To meet the convenience property, we offer a multimodal biometric system based on face and voice biometrics. The choice of those two traits arises from the fact that, both of them are non intrusive, convert, easy to capture, most commonly used in human machine interface and socially accepted.
To solve the accuracy problem, we have studied two seperate systems namely personal identification and personal verification systems. In each system, each biometric modality was studied separately. Then fusion was performed using Gaussian mixture model GMM and neural network NNT techniques at various levels. As NNT fusion gave us the best results a study of NNT implementation on FPGA was processed. through this work, both user-dependent and user-independent processing have been examined. We have proved that, user dependent processing not oly enhanced the system accuracy but also gave us the ability to simplify the identification fusion process identification fusion was performed by considering both the match scores that decision of underlaying system.
To make system scalable speaker pruning was performed using gender classification to filter the biometric database, which in turn, will improve the system speed and accuracy.
Finally to meet the security challenge, this work offers a reliable identity management system by applying several ways to resist external attacks on biometric systems