Enhancement of E-Banking System in Iraq by web application-based authentication system using face recognition

Abstract

Enabling customers to access their bank accounts electronically brings many benefits, facilitating interaction between customers and banks, and saving time and effort. Customers no longer need to visit physical bank branches or exchange manual transactions. This transformation can be achieved through a strong electronic authentication system that allows customers to access their bank accounts without visiting a bank branch. The foundation of this system is a powerful and unique biometric authentication feature – facial recognition. Have developed a web application that uses real-time facial recognition for electronic banking systems. Iraqi banks previously lacked such advanced methods of electronic authentication, and research seeks to fill this gap. Implementing this system increases customer confidence regarding the security of their accounts. By using a real-time web application that processes real data, the system can accurately identify and tag individuals, and grant or deny them access as appropriate. Two web applications were created, the first using Olivette's global database for image matching and recognition. The second real-time application allows users to access their bank accounts by taking instant photos of them, and detecting, processing, and identifying the people in the photos. If an unauthorized person gains access, access will be denied. Using the Principal Component Analysis (PCA) algorithm in conjunction with the Eigenface technique and the Support Vector Machine (SVM) classifier, it is possible to achieve high accuracy in face recognition in both web programs. Specifically, able to achieve an accuracy of up to 92% and then improve it to 94%.

Keywords

• E-banking system, Authentication, face recognition, web application, PCA, and SVM.