A proposed algorithms to design support multimodal biometric system

Abstract EN

This paper is an attempt to address the biometric security issue and improve the system accuracy through introducing a design for multimodal biometric verification system using multiple traits (Iris, fingerprint) and adding another phase called liveness detection to the phases of multimodal system the purpose of this phase is to protect the multimodal biometric systems against spoofing attacks.

The system is tested in two levels, unimodal level and multimodal level (fusion level).

in unimodal level two tests have been performed, one for iris verification phase performed on two types of database MMU DB (Multi Media University database) for 180 samples and CASIA DB (Chinese Academy of Sciences database) for 90 samples.

and gave accuracy (99.44%) with FAR (False Acceptance Rate) of (0.0277) and FRR (False Reject Rate) (0.0055) for MMU DB, and accuracy (97.77%) with FAR of (0.0333) and FRR (0.0222) for CASIA DB, and other for fingerprint verification phase performed on database collected from two types of database for 60 samples and gives accuracy of 95% with FAR of 0.1% and FRR of 0.05%.

In multimodal level the system is tested on database composed of 60 samples for iris images and 60 samples for fingerprint images and gives an overall accuracy of 100% with FRR of 0%, and FAR of 0.0166%.