A Proposed Model of Advanced Security System in ATM: Implementation of Face Recognition and Finger Print Recognition

Abstract

ATM is an essential part of every human life as it enables the customer of a bank to perform financial transactions at any time and without the direct interaction with the bank employees. But, with the increase in the use of ATM, the machines are being prone to hacker attacks, fraud, robberies and security breaches. False keypad, hidden cameras are being used to get the account information and PIN of a user which lead to card cloning or skimming. In this paper, the authors introduce a fast and robust method for the advancement of ATM security which applies a combination of face recognition and fingerprint recognition system as biometric customer authentication and thereby obsoletes the use of PIN. In this method, after the user inserts the card in the card reader, his/her face undergoes the detection and matching process using PCA face detection and recognition algorithm with Eigenfaces method. If a match occurs, the user’s fingerprint is scanned and matched using Minutiae based matching algorithm. If the result comes positive, the user is allowed to process the usual transaction. The experimental results carried out with a laptop webcam and a commercial fingerprint scanner, illustrates the robustness and efficiency of the approach even in low light environments. The challenges and limitations faced during the research and some suggestions for future research directions are also discussed in this paper.