Feature Level Fusion of Iris and Voice.

Sameh, Sally Mohamed; Moustafa, Hossam El-Din Salah; Zaki, Fayez Wanis

doi:10.21608/bfemu.2020.96269

Feature Level Fusion of Iris and Voice.

Document Type : Research Studies

Authors

¹ Electronics and Communications Engineering Department , Faculty of Engineering, Mansoura University, Mansoura, Egypt

² Electronics and Communications Engineering Department, Faculty of Engineering, Mansoura University, Mansoura, Egypt

10.21608/bfemu.2020.96269

Abstract

Biometric systems are vastly used by various organizations for different security applications. The main use of such systems is in authentication and identification applications which including computer network login, passport control, corpse identification, electronic data security, terrorist identification, Internet access. Unimodal biometric systems suffer from a lot of problems like high error rate, low performance and imposters’ attack so the demand of Multimodal biometric systems takes place. Multimodal biometric systems have several advantages over unimodal biometric systems such as non-universality, larger population coverage, lower error rates, higher performance and higher genuine acceptance rates. In this paper, a study of multimodal fusion of voice and Iris at feature level is presented. The features are extracted from the voice signals using Power-Normalized Cepstral Coefficients (PNCC) and from the preprocessed iris images using Single Value Decomposition (SVD). The experiment have been done using samples collected from faculty of Engineering , Mansoura university for voice and CASIA iris database for iris which gave accuracy of 98.4% after fusion. This result is acceptable and gives high accuracy than using voice and iris individually.

Keywords

Main Subjects