Design and Hardware Implementationof a Secure Bi-Modal Biometric System
| Autor: | Jia-hua Lu, 呂佳樺 |
|---|---|
| Rok vydání: | 2006 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 94 In recent years, biometrics has become an important topic for public safety and information security. This technique relies on human appearances and behavioral characteristics to identify or verify a person’s identity. However, the performance of single modal biometric system is usually degraded due to the environmental influence or the physiological defect. In this thesis, we design a bi-modal biometric system based on face/iris biometric decision fusion in view of improving substantially the recognition performance and the security in a biometric authentication system. In the proposed method, the feature vectors of face modal and iris modal are verified respectively by two independent PNN neural networks. The inference probabilities of the PNN are integrated into a fuzzy neural network classifier which will output the final verification decision. Our experiment reveals an excellent performance by the bi-modal biometrics (EER=0.823%), which is highly superior to the face biometrics (EER=2.62%) and the iris biometrics (EER=2.21%). We implement finally the proposed bi-modal biometric system on a FPGA single chip by means of the high-level hardware synthesis. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|