Secure authentication for mobile banking using facial recognition

Abstract

THE CURRENT SYSTEM OF MOBILE BANKING IN NIGERIA CONSISTS OF TWO WAY AUTHENTICATION (USERNAME AND PASSWORD) WHICH CAN BE FORGOTTEN OR STOLEN. THIS PAPER PROPOSES A SECURE AUTHENTICATION FOR MOBILE BANKING USING FACIAL RECOGNITION TO THE EFFECT OF IMPROVING ON THE EXISTING SYSTEM, AND THEREBY SUPPORTING THE ACTUALIZATION OF CASHLESS SOCIETY. AN OVERVIEW AND LIMITATIONS OF THE CURRENT SYSTEM ARE PRESENTED. THE HIGH LEVEL DESIGNS OF THE PROPOSED SYSTEM ARE THEN PRESENTED. THE SYSTEM IS THEN SIMULATED USING JAVA PROGRAMMING LANGUAGE AND TESTED USING SIMULATED DATABASES OF NIGERIA COMMUNICATION COMMISSIONS (NCC) AND THE FACILITATING BANK. THE SYSTEM WAS FOUND TO PERFORM WITH A MAXIMUM RESPONSE TIME OF SEVEN MINUTES, AND FALSE ACCEPTANCE RATE (FAR) OF 3%. COMBINING THIS SYSTEM WITH ONE OR MORE OTHER FORMS OF BIOMETRIC TECHNOLOGIES SUCH AS FINGER VEIN, IRIS AMONG OTHERS WILL NO DOUBT GIVE A FRAUD PROOF PLATFORM FOR MOBILE PHONE BANKING.